bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
1 The genetic architecture underlying ecological adaptation in
2 Atlantic herring is not consistent with the infinitesimal model
3
4 Fan Han1, Minal Jamsandekar2, Mats E. Pettersson1, Leyi Su1, Angela Fuentes-Pardo1, Brian
5 Davis2, Dorte Bekkevold3, Florian Berg4,5, Michele Cassini6,7, Geir Dahle5, Edward D.
6 Farell8, Arild Folkvord4,5, Leif Andersson1,2,9,*
7
8 1Department of Medical Biochemistry and Microbiology, Uppsala University, Uppsala,
9 Sweden.
10 2Department of Veterinary Integrative Biosciences, Texas A&M University, College Station,
11 USA.
12 3National Institute of Aquatic Resources, Technical University of Denmark.
13 4Department of Biological Sciences, University of Bergen, Bergen, Norway.
14 5Institute of Marine Research, Bergen, Norway.
15 6Department of Aquatic Resources, Institute of Marine Research, Swedish University of
16 Agricultural Sciences, Lysekil, Sweden.
17 7Department of Biological, Geological and Environmental Sciences, University of Bologna,
18 Bologna, Italy.8School of Biology and Environmental Science, Science Centre West,
19 University College Dublin, Belfield, Dublin 4, Ireland.9Department of Animal Breeding and
20 Genetics, Swedish University of Agricultural Sciences, Uppsala, Sweden.
21 *Corresponding author: Email:[email protected]
22
1 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
23 Abstract
24 Atlantic herring is widespread in North Atlantic and adjacent waters and is one of the
25 most abundant vertebrates on earth. This species is well suited to explore genetic
26 adaptation due to minute genetic differentiation at selectively neutral loci. Here we
27 report hundreds of loci underlying ecological adaptation to different geographic areas
28 and spawning conditions. Four of these represent megabase inversions confirmed by
29 long read sequencing. The genetic architecture underlying ecological adaptation in the
30 herring is in conflict with the infinitesimal model for complex traits because of the large
31 shifts in allele frequencies at hundreds of loci under selection.
32
33 Main
34 Atlantic herring (Clupea harengus) constitutes the basis for one of the world’s most important
35 commercial fisheries1 and has been a valuable food resource throughout human history in
36 North Europe. It is one of the most abundant vertebrates on earth with a total estimated
37 breeding stock of one trillion individuals2. Its short-term (<10,000 years) effective population
38 size is enormous whereas the long-term effective population size is much more restricted
39 most likely due to the impact of periods of glaciation3. As a consequence, the nucleotide
40 diversity is moderate (~0.3%)3, only three-fold higher than in human although the census
41 population size is at least two orders of magnitude higher in herring and until recently (>1,000
42 years ago) many orders of magnitude higher.
43 Detailed population genomic studies of Atlantic herring are justified for two reasons.
44 Firstly, a better understanding of the population structure and development of diagnostic
45 markers have the potential to revolutionize the procedures of stock assessments of herring that
46 has been in operation for more than hundred years, which will add key scientific knowledge
47 to safeguard a sustainable fishery. Secondly, the Atlantic herring can be used as a model to
2 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
48 study the genetic architecture underlying phenotypic diversity and ecological adaptation. The
49 herring has adapted to different marine environments including the brackish Baltic Sea, where
50 salinity drops to 2-3 PSU and water temperature shows a much larger variation among
51 seasons than in the Atlantic Ocean with typically 34-35 PSU (Extended Data Fig. 1).
52 Furthermore, herring spawn during different periods of the year, which involves
53 photoperiodic regulation of reproduction. There is minute genetic differentiation at selectively
54 neutral loci among herring populations from different geographic regions3,4 and the
55 distribution of per locus FST values deviates significantly from the one expected for
56 selectively neutral loci with an excessive number of outlier loci demonstrating signatures of
57 selection3,5. These properties make the Atlantic herring a powerful model to explore how
58 natural selection shapes the genome in a species where genetic drift plays a minor role in
59 genome-wide population divergence.
60 The genetic basis for complex traits and disorders is of central importance in current
61 biology and human medicine. The aim of the present study was to test whether the genetic
62 architecture underlying ecological adaptation in herring is consistent with the expectations
63 from Fisher’s classical infinitesimal model for polygenic traits6. This has been accomplished
64 by comprehensive analysis of whole genome sequence data from 53 population samples
65 spread across the entire species distribution.
66
67 Results
68 Genome sequencing
69 We performed pooled whole-genome sequencing using 28 population samples of Atlantic
70 herring. These were analyzed together previously published data3,7, making a total number of
71 53 population samples. The current study has added populations around Ireland and Britain,
72 from Norwegian fjords and from the Southern Baltic Sea. Together the 53 population samples
3 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
73 cover the entire species distribution and include spring, autumn, winter and summer spawning
74 stocks (Fig. 1 and Table S1). A previously sequenced population of Pacific herring (Clupea
75 pallasii) was used as a closely related outgroup in the bioinformatic analysis. We pooled
76 genomic DNA from 35-110 individuals per population, and conducted whole-genome
77 resequencing to a minimum of 30 coverage per population. In order to generate individual
78 genotype data we sequenced 12 fish from four localities to ~10 coverage per individual.
79 Together with data from previous studies3,4,7, we used a total of 55 individually sequenced
80 herring in the analysis (Table S2). All sequence reads were aligned to the recently released
81 chromosome-level assembly for Atlantic herring8, and sequence variants were called and
82 filtered with a stringent pipeline (Methods). In total, we identified ~11.5 million polymorphic
83 biallelic sites among the 53 population samples and ~15.9 million sites when including
84 Pacific herring.
85
86 Loci under selection reveal population structure
87 A genetic distance tree, based on genome-wide SNPs, groups the 53 population samples into
88 seven primary clusters: (i) autumn- and (ii) spring-spawning herring from the brackish Baltic
89 Sea, (iii) populations from the transition zone, close to the entrance to the Baltic Sea,
90 spawning at lower salinity than in the Atlantic Ocean, (iv) Norwegian fjord populations, (v)
91 populations around Ireland and Britain, (vi) autumn- and (vii) spring-spawning herring from
92 the North Atlantic Ocean (Extended Data Fig. 2). This clustering, in general, fits with their
93 geographical origin, but the branch lengths separating some subpopulations are very short,
94 e.g. autumn- and spring-spawning populations from the North Atlantic Ocean. Pairwise FST
95 values among all populations are in the range 0.013 to 0.061 (Fig. S1).
96 To explore the contribution of loci under selection to the observed population
97 structure, we separated SNPs showing no significant genetic differentiation between
4 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
98 populations from those showing strong genetic differentiation based on their standard
99 deviation of allele frequencies across Atlantic and Baltic populations (Methods; Fig. S2). We
100 carried out Principal Component Analysis (PCA) based on these two sets of markers
101 independently. The PCA using 169,394 undifferentiated markers separated some of the major
102 groups but three distinct groups (Atlantic Ocean spring and autumn spawners, and
103 populations around Ireland and Britain) were merged into one tight cluster (Fig. 2a). The first
104 two principal components explained only ~11% of the variance. This is in line with the
105 conclusion that the majority of the genome shows minute genetic differentiation across the
106 entire species distribution of Atlantic herring3,4,7. In contrast, the PCA based on only 794
107 markers showing the most striking genetic differentiation separated the populations into the
108 seven main clusters (Fig. 2b), consistent with the genetic distance tree (Extended Data Fig. 2).
109 The first principal component explained 43% of the variance. These two patterns observed for
110 the two sets of markers suggest that the observed population structure in herring can be
111 accounted for by a small fraction of genetic markers associated with loci underlying
112 ecological adaptation.
113 The PCA analysis based on the differentiated SNPs revealed some intriguing results as
114 regards population classifications (Fig. 2b). Two population samples (#24 and 25) both from
115 Landvikvannet in southern Norway clustered more closely with populations from the
116 transition zone than with local fjord populations from western Norway. Landvikvannet is a
117 unique brackish lake created in 1877 by opening a canal between the ocean and a freshwater
118 lake, now harboring a local population of herring adapted to this brackish environment9. The
119 populations around Ireland and Britain showed a relatively high degree of diversity, despite
120 their geographical proximity. Populations within this cluster do not show such distinct sub-
121 clusters according to spawning season as they do for instance in the Baltic Sea. One of the
122 populations from Schlei (Sample 21), collected as autumn spawning, clustered with spring-
5 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
123 spawning populations. Interestingly, examinations of the otoliths from fish in this sample,
124 using the method by Clausen et al.10, showed that they hatched in spring. Thus, this must
125 represent a group of herring that has switched spawning season from spring to autumn, maybe
126 because their nutritional status was insufficient to support spawning11, demonstrating that
127 spawning time in herring is controlled by a combination of genetic and environmental
128 factors12,13.
129 Taken together, in Atlantic herring, genetic markers associated with loci under
130 selection provide a much better resolution to distinguish population structure than random
131 neutral genetic markers.
132
133 Genetic signatures of ecological adaptation
134 To explore the genetic architecture underlying ecological adaptation in the herring, we formed
135 superpools of populations representing the major groups detected using the PCA analysis
136 (Fig. 2b), and performed four genome-wide contrasts (Table S3), (i) Baltic herring vs.
137 Northeast Atlantic herring, (ii) spring-spawning vs. autumn-spawning herring, (iii) Atlantic
138 herring from the waters around Ireland and Britain vs. other parts of the Northeast Atlantic
139 Ocean, and (iv) Atlantic herring from the North-Atlantic Ocean vs. all other Atlantic herring
140 samples but excluding herring from the Baltic Sea and Norwegian fjords (Fig. 2b). In each
141 contrast, we compared allele frequency differentiation on a per SNP basis using a χ2 test and
142 used a stringent significance threshold (P <1x10-10, Methods).
143 Baltic herring vs. Northeast Atlantic herring. We divided this contrast into two
144 replicates using the genetically distinct spring- and autumn-spawning populations from each
145 region. The Baltic/Northeast Atlantic contrast using 19 spring-spawning and 9 autumn-
146 spawning populations resulted in a total number of 115 and 39 loci, respectively, reaching
147 statistical significance (Fig. 3a). As many as 30 of these loci reached significance in both
6 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
148 comparisons (highlighted in red in Fig. 3a). These loci are the best candidates for adaptation
149 to the environmental conditions in the brackish Baltic Sea irrespective of spawning time. A
150 prime example of such a locus is LRRC8C (leucine-rich repeat-containing 8 protein, subunit
151 C) (P < 10-139, in spring-spawners and P < 10-62 in autumn-spawners) which encodes a
152 volume-regulated anion-channel. Another locus that was highly significant in both replicate
153 comparisons is PRLR (prolactin receptor) on Chr12 highlighted in Fig. 3b. The most
154 significant association at this locus is a cluster of non-coding SNPs located in a gene desert
155 upstream of PRLR (Extended Data Fig. 3). This hormone-receptor has an important role
156 during lactation in mammals and a critical role for osmoregulation and adaptation to salinity
157 in fish14. The loci that do not replicate in the two subsets are not false positives but are less
158 likely to be directly related to adaptation to brackish water. A striking example of such a locus
159 is the large 7.8 Mb region on Chr12 that shows no genetic differentiation in the spring-
160 spawning contrast, but very strong genetic differentiation between autumn-spawning Atlantic
161 and Baltic herring (Fig. 3b). This region corresponds to a recently reported inversion8 (see
162 further data below).
163 Spring-spawning vs. autumn-spawning herring. We divided this contrast into two
164 replicates: (i) populations from Baltic Sea and (ii) populations from the NortheastAtlantic
165 Ocean. The spring-/autumn-spawning contrast involved 16 Baltic and 12 Northeast Atlantic
166 populations, which yielded 31 and 13 significant loci, respectively (Fig. 3c). Only 7 loci were
167 significant in both replicates. Out of the 7 replicated loci, 3 were found within 5 Mb on Chr15
168 (Fig. 3d). The most significant signal among those contains the TSHR (Thyroid Stimulating
169 Hormone Receptor) locus (P < 10-106 in Baltic populations and P < 10-82 in Atlantic
170 populations), which has a well-defined role in the pathway regulating seasonal reproduction
171 in vertebrates15,16 and CALM1B (Calmodulin), which regulates gonadotrophin-releasing
172 hormone signaling via the MAPK pathway17. Another significant locus on Chr15 is SOX11B
7 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
173 (SRY-Box Transcription Factor 11), which plays an essential role in sex determination in
174 zebrafish18. Notably, the most significant sequence variant at this locus (P < 10-101 in Baltic
175 populations and P < 10-88 in Atlantic populations) is a single-base change in a non-coding
176 region located 39 kb upstream of SOX11B (Extended Data Fig. 3b), suggesting that it may be
177 a causal regulatory mutation. The third highly significant locus on Chr15 contains ESR2A
178 (Estrogen Receptor Beta) and SYNE2 (Spectrin Repeat Containing Nuclear Envelope 2), and
179 these two genes are only 100 kb apart (Fig. 3d). In fact, a closer examination of this region
180 shows that there are two distinct signals at the two genes. ESR2A is an obvious candidate gene
181 for a reproduction-related phenotype, suggesting that the signal at SYNE2 may reflect
182 regulatory variants affecting ESR2A expression. However, SYNE2 encodes an important
183 structural protein and a premature stop codon in this gene is causing defects in photoreceptors
184 in mice19, suggesting that it may in fact has a role in photoperiod regulation of reproduction in
185 herring.
186 In addition to the signals on Chr15, we also found other loci that are associated with
187 adaptation to different spawning seasons, such multiple members of MYHC (myosin heavy
188 chain) gene family on Chr12, which may be related to a reported association between
189 myogenesis and plasticity of seasonal development in herring20. Another locus is HERPUD2
190 (Homocysteine Inducible ER Protein with Ubiquitin like Domain 2) on Chr19, a gene that has
191 no known function in reproductive biology.
192 Atlantic herring from the waters around Ireland and Britain show distinct
193 signatures of selection. An important geographical area missing in our previous genetic
194 screens are the waters around Ireland and Britain, which are inhabited by the southernmost
195 ecomorphs of herring in the Northeast Atlantic that spawn in warmer sea water than other
196 herring populations (Extended Data Fig. 1). We grouped 10 populations from this region as
197 one superpool, and compared its genome-wide allele frequency with another superpool, which
8 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
198 consisted of 4 neighboring populations from the Northeast Atlantic Ocean (Fig. 4a).
199 Following the definition of independent loci (see Methods), we found only 12 independent
200 loci showing significant genetic differentiation in this contrast. Several of these are the loci
201 also detected in the contrast between spring- vs autumn-spawning herring (TSHR, HERPUD2,
202 etc.), as expected due to the difference in proportion of autumn-spawners (40% vs. 0%).
203 However, four of the major loci on Chr6, 12, 17 and 23 span exceptionally large regions on
204 four different chromosomes (Fig. 4a). Unlike a typical selective sweep where divergence of
205 the markers forms a bell shape, the zoom-in profile of the significant SNPs at each locus
206 formed a block-like pattern, where a sharp change of divergence was observed at borders
207 (Fig. 4b-e). The block-like pattern, suggesting suppressed recombination, and the sharp
208 borders of the regions are consistent with the presence of inversions or other structural
209 rearrangements.
210 Further exploration of allele frequencies at diagnostic markers at each putative
211 structural variant showed that the majority of the populations from Ireland and Britain tend to
212 be homozygous for S (South) haplotypes while N (North) haplotypes dominate in all other
213 populations (Extended Data Figs 4 and 5). At the Chr6 locus, all other Baltic and Atlantic
214 populations tend to be fixed for the N haplotype, except a few populations from Norwegian
215 fjords and the transition zone, which suggests that Chr6 S haplotype is contributing to
216 adaptation to local ecological conditions around Ireland and Britain, and occurs at a
217 particularly high frequency in populations spawning near the Isle of Man in the Irish Sea, the
218 Downs near the English Channel and in the Celtic Sea (Extended Data Fig. 6a). The most
219 striking differentiation was observed on Chr12, where the -log10(P) value exceeded 80 (Fig.
220 4c). In fact, it coincides with a recently reported inversion, which is potentially associated
221 with adaptation to water temperature at spawning8. The haplotype frequencies among
222 populations supports this hypothesis (Extended Data Fig. 6b). The individual sequence data
9 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
223 reveal that recombinant versions of this inversion are frequent in the Baltic Sea, but the exact
224 breakpoints of the recombined regions within the inversion are variable (Extended Data Fig.
225 5).
226 The putative inversion on Chr17 shows a similar pattern as the Chr6 polymorphism in
227 the southernmost populations in East Atlantic where the S haplotype is predominant, but some
228 populations near the transition zone also carry a large proportion of the S haplotype (Extended
229 Data Fig. 6c). No population is fixed for the S haplotype at the Chr23 locus, although S is still
230 the dominant haplotype in the majority of the southernmost populations (Extended Data Fig.
231 6d). The association between the haplotype frequencies and latitude across populations for the
232 Chr12, Chr17 and Chr23 variants may underlie adaptation to ecological factors that are
233 related to water temperature (see Extended Data Fig. 1). The Chr6 locus seems to be
234 associated with ecological adaptation specifically to habitats around Ireland and Britain.
235 We constructed a neighbor-joining tree based on all SNPs from each of the four loci
236 and estimated the nucleotide diversity of each haplotype. We found that, in general, both N
237 and S haplotypes at each locus show lower nucleotide diversity in the range 0.11% to 0.31%
238 than the genome average (~0.3%) and S haplotypes tend to be more variable than N
239 haplotypes (Extended Data Fig. 7). This suggests that genetic recombination has been
240 restricted between haplotypes and that N haplotypes are younger than the S haplotypes or
241 have been more affected by population bottlenecks, possibly before or during the last
242 glaciation.
243 Signatures of selection in herring from the North Atlantic Ocean. Another pattern
244 was that populations from the North Atlantic Ocean formed two major groups (Fig. 2b;
245 Extended Data Fig. 2). These groups include spring- and autumn-spawning populations from
246 Nova Scotia, Newfoundland, Greenland, Iceland and those representing the Norwegian
247 Spring Spawning (NSS) herring (samples #41 and 44). The latter is the most abundant herring
10 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
248 stock in the world with a current adult breeding stock of >2x1011 individuals21. The stock
249 spawns along the Norwegian coast in early spring, followed by a northward larval drift and
250 later feeding migration into the Norwegian Sea. After feeding, the majority of the stock
251 overwinters in the fjords of northern Norway22. We designed a contrast between this group of
252 10 population samples and all other 21 population samples but excluded those from the Baltic
253 Sea and Norwegian fjords. The whole-genome screen revealed, as expected, the four major
254 loci on Chr6, 12, 17 and 23, that were significant in the partially overlapping former contrast
255 (Fig. 5a). However, the most striking signal of selection in the contrast involving the North
256 Atlantic herring is a locus on Chr2 overlapping a single gene: the Aryl Hydrocarbon Receptor
257 (AHR) (Fig. 5a). Four of the ten top SNPs at this locus are missense mutations, including the
258 one (Ser346Thr) that by far showed the strongest genetic differentiation (P < 10-87; Fig. 5b),
259 suggesting that it is likely to be a causal variant. The non-reference allele Thr346 occurs at a
260 high frequency in all population samples from the North-Atlantic Ocean whereas the
261 reference allele Ser346 dominates in all other population samples (Extended Data Fig. 8a).
262 However, the Thr346 allele is widespread and there is a tendency for a North-South cline both
263 in the Baltic Sea and around Ireland and Britain with a slightly higher frequency of Thr346 in
264 the North.
265 Besides the AHR locus, we noted several other highly significant loci not identified in
266 other contrasts. The most significant SNP on Chr7 (P < 10-47; Fig. 5a) is located ~1kb
267 upstream of solute carrier family 12 (sodium/chloride transporter) member 2
268 (SLA12A2/NKCC1), a gene that shows altered expression in response to changes in salinity in
269 spawning sea lamprey23. Another interesting locus is THRB (thyroid hormone receptor beta)
270 on Chr19 encoding a medically important nuclear receptor. Mutations in THRB cause
271 generalized thyroid hormone resistance in human24 and a recent CRISPR study in zebrafish
272 revealed its function in photoreceptor development25. At the herring THRB locus, all
11 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
273 populations from high salinity conditions (34-35 PSU) are fixed for an Atlantic haplotype
274 whereas, with a few exceptions, a “Baltic” haplotype, which must have been introgressed
275 from the sister species Pacific herring, dominates in all populations from the Baltic Sea, the
276 transition zone and Norwegian fjords (Extended Data Fig. 9). The only sequence difference
277 between this introgressed haplotype and the haplotype present in Pacific herring from
278 Vancouver is a missense mutation in THRB, Gln40His.
279
280 Identification and characterization of inversions using PacBio long reads
281 PacBio long read sequencing of size selected fragments (>40 kb) from one male individual
282 from the Celtic Sea was used in an attempt to validate candidate inversions on Chr6, 12, 17
283 and 23 (Fig. 4). This individual was not informative for the previously documented inversion
284 on Chr128 because it was homozygous for the S haplotype present in the reference assembly.
285 PacBio sequencing provided conclusive evidence for the presence of inversions on Chr6 and
286 17 (Fig. 6). On Chr6, the distal breakpoint is at 24,868,581 bp and the proximal breakpoint
287 lies between 22,280,036 bp and 22,290,000 bp. Its exact position is difficult to deduce due to
288 complexity of the region and because the reference assembly is not complete at the proximal
289 breakpoint. A representative PacBio read spanning the proximal breakpoint on Chr6 is shown
290 in Fig. 6a where the read starts within the inversion in the orientation towards the distal
291 breakpoint, continues across the proximal breakpoint, and ends outside the inversion. This
292 pattern is supported by 34 reads. Fig. 6b illustrates the Chr17 inversion where a PacBio read
293 maps before the proximal breakpoint and continues to the sequence proximal to the distal
294 breakpoint in the reference assembly. There are 36 reads supporting this pattern with
295 25,805,445 bp as the proximal breakpoint and 27,568,511 bp as the distal breakpoint. The
296 inserted dotplot shows the presence of a 5 kb duplicated sequence in the reference N
297 haplotype, located with an opposite orientation at the two breakpoints. This 5 kb sequence is
12 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
298 present as a single copy in the S haplotype of the PacBio individual, implying that the N
299 haplotype is most likely the derived state, and that the inversion and duplication occurred
300 during the same meiotic event. This interpretation is consistent with the higher nucleotide
301 diversity among S haplotypes (0.31% vs. 0.24%, Extended data Fig. 7c). There are a total of
302 44 and 66 genes present within the inverted regions of Chr6 and 17, respectively. The
303 inversion breakpoints do not disrupt any coding sequence, but the proximal breakpoint of the
304 Chr6 inversion is located very close to exon 1 of the methyl transferase 3 (PRMT3) gene
305 suggesting that the inversion might affect its expression pattern. It is a highly conserved gene
306 in vertebrates and involved in bone development in human26. Complete lists of genes within
307 the inverted regions and their flanking sequence are provided in Table S4.
308 The PacBio data only provided suggestive evidence for an inversion on Chr23. A
309 proximal inversion breakpoint was indicated at 16,378,067 bp, but the characterization of the
310 distal breakpoint was hampered because the individual used for the reference assembly8 is
311 most likely heterozygous for this inversion.
312
313 Discussion
314 The first population genetic studies on Atlantic herring in the early 1980s were based on a
315 dozen allozyme markers and revealed the unexpected result of no significant allele frequency
316 differences even between Atlantic and Baltic herring considered as subspecies27,28.
317 Subsequent studies using microsatellites and small number of SNPs largely confirmed this
318 observation but with occasional genetic markers showing genetic differentiation5,29. Whole
319 genome sequencing has completely changed the picture, we now find striking genetic
320 differentiation but at a limited number of loci (Figs 3-5). Our data are consistent with a model
321 where the Atlantic herring has a genomic “tool box” of a few hundred major loci associated
322 with ecological adaptation to local climate conditions such as water temperature, salinity,
13 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
323 light conditions, and preferred spawning time. We propose that this genetic adaptation is
324 related to life history parameters such as migratory behavior, timing of spawning and
325 ecological conditions during larval development. The Atlantic herring has a high fecundity
326 where a female can release tens of thousands of eggs at spawning. A perfect match between
327 genotype and environmental conditions gives potential for a highly successful reproduction.
328 An important practical implication of the current study is that sequence variants
329 associated with ecological adaptation best reveal the population structure of the Atlantic
330 herring (Fig. 2b), whereas neutral markers are less useful for population identification (Fig.
331 2a). A PCA analysis based on only about 800 SNPs out of the more than 10 million SNPs
332 detected in the study distinguish all major groups of Atlantic herring and reveal allele
333 frequency differences between subpopulations within major groups. For instance, the
334 Ser346Thr mutation in AHR alone is able to efficiently distinguish the North Atlantic Ocean
335 groups from all other groups of herring (Extended Data Fig. 8a). The current study has
336 established the diagnostic SNP markers that may be used to improve the collation of data for
337 fisheries stock assessment. This is necessary in order to identify and address the mismatches
338 between biological populations and management units or stocks, which can hamper the
339 development and implementation of effective management of this important natural
340 resource30,31.
341 The genetic architecture underlying multifactorial traits and disorders has been
342 intensively studied in recent years, in particular based on human genome-wide association
343 studies (GWAS), and the general finding is that most traits are highly polygenic with large
344 number of loci each explaining a tiny fraction of the genetic variance32. This is usually
345 referred to as consistent with a quasi-infinitesimal model, because the data largely conform
346 with the predictions of Fisher’s classical infinitesimal model6 which constituted the
347 foundation for quantitative genetics theory. A prime example of a trait consistent with the
14 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
348 quasi-infinitesimal model is stature in human. A study based on more than 250,000
349 individuals revealed more than 400 loci affecting this trait but these loci did not explain more
350 than 20% of the heritability33, implying that it must be more than thousand loci affecting
351 stature in humans. The genetic architecture underlying ecological adaptation in Atlantic
352 herring is not consistent with the quasi-infinitesimal model. Ecological adaptation in herring
353 is polygenic but the number of loci showing strong genetic differentiation is in the hundreds
354 and not in the thousands. However, it is possible that we only see the tip of the iceberg and
355 there may be a large number of loci with tiny effects that also affect ecological adaptation.
356 The herring data are nevertheless a flagrant deviation from the infinitesimal model because
357 this theory predicts that a response to natural selection should only cause minor changes in
358 allele frequencies34. This is because each individual locus only explains a tiny fraction of the
359 genetic variance and should not result in major shifts in allele frequencies, such as those
360 reported here. An important difference between human and herring data, is that the signatures
361 of selection in herring reflect evolutionary processes that may have happened over many
362 generations. For instance, the herring has adapted to the brackish Baltic Sea after it was
363 formed about 10,000 years ago35. In contrast, the human GWAS data reflect standing genetic
364 variation, which includes many slightly deleterious variants, the majority of these will never
365 contribute to a selection response for a fitness trait.
366 The genetic architecture underlying ecological adaptation detected here makes perfect
367 sense in light of the ecological challenges involved. The most dramatic differences in biotic
368 and abiotic factors occur between the marine Atlantic Ocean and the brackish Baltic Sea
369 which differ in regard to salinity, variation in water temperature over the year (Extended Data
370 Fig. 1), light conditions36, plankton production and presence of predators. In this contrast, we
371 document hundreds of loci showing significant genetic differentiation. In other contrasts
15 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
372 between geographically distinct ecoregions, the numbers of detected loci are much fewer, in
373 the range 10-30 (Figs 3-5).
374 It has been exceedingly challenging to move from GWAS peaks to the identification
375 of causal variants in human and other species, despite enormous efforts. The reason is that
376 strong linkage disequilibrium among multiple sequence variants showing an equally strong
377 association to phenotype makes it difficult or impossible to provide genetic evidence for
378 causality. In contrast, the genome-wide screens in Atlantic herring immediately reveal strong
379 suggestive evidence of causality at some loci. A striking example concerns rhodopsin (RHO),
380 showing strong genetic differentiation between Atlantic and Baltic herring (Fig. 3a). A
381 missense mutation in this gene (Phe261Tyr) has a critical role for adaptation to the red-shifted
382 light conditions in the Baltic Sea36. Interestingly, about one third of all fish adapted to red-
383 shifted light conditions in freshwater and brackish water carry Tyr261 like the Baltic herring,
384 an amazing example of convergent evolution36. The present study revealed at least two
385 additional examples of outstanding candidates for being causal mutations, the non-coding
386 single base change upstream of SOX11B (Extended Data Fig. 3b) and the Ser346Thr missense
387 mutation at AHR (Fig. 5b). These are candidate causal mutations based on far stronger
388 statistical support than any other sequence variant in the corresponding regions. How can
389 genomic regions showing signals of selection in these cases be so narrow? There are two
390 possible explanations, either it is because these are “old” polymorphisms that have segregated
391 for such a long time that linkage disequilibrium to flanking markers has eroded leaving the
392 causal variant in splendid isolation, or alternatively the same causal variant has occurred by
393 mutation multiple times on different haplotypes. Although the mutation rate in the Atlantic
394 herring is exceptionally low2, in each generation every nucleotide site in the herring genome
395 is hit by many mutations due to the enormous population size. Thus, herring is not a species
396 where evolution is restricted by lack of genetic variation and therefore has to rely on standing
16 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
397 genetic variation. The observation of sequence variants with outstanding strength of
398 associations, implying substantial functional effects, is a deviation from the infinitesimal
399 model.
400 A unique observation in this study is the Ser346Thr missense mutation in the aryl
401 hydrocarbon receptor (AHR) gene. Thr346 must be the derived allele since all tested Pacific
402 herring were homozygous Ser346. Furthermore, residue 346 is located in the highly
403 conserved PAS B domain and Ser346 is highly conserved among vertebrates (Extended Data
404 Fig. 8b), suggesting a functional significance of Ser346Thr. AHR belongs to the basic helix-
405 loop-PER-ARNT-SIM (bHLH-PAS) family of transcription factors. AHR is cytosolic signal
406 sensor that binds planar aromatic hydrocarbons and is then translocated to the nucleus where
407 it forms a heterodimer with the AHR nuclear translocator (ARNT) and affect
408 transcription37,38. The endogenous ligands of AHR include tryptophan-derived metabolites
409 and dietary indoles, while the most well-known exogenous high-affinity ligand is dioxin37. In
410 human medicine, AHR function is of relevance for toxicity response, inflammation,
411 autoimmune disorders, cancer, circadian clock and for sensing bacterial virulence and thereby
412 activate antibacterial responses and control sepsis37,38. It has recently been reported that a
413 population of killifish has been able to colonize highly polluted waters due to the presence of
414 a deletion knocking out AHR function39. In herring, the derived Thr346 allele occurs at a high
415 frequency in the least polluted waters in the North Atlantic Ocean, e.g. around Greenland and
416 Iceland, suggesting that it is unlikely that this adaptation is related to the presence of
417 exogenous metabolites. A serine to threonine change is a conservative amino acid substitution
418 since they are both polar, neutral amino acids with a hydroxyl or methyl side chain,
419 respectively. This suggests that this missense mutation fine-tunes rather than disrupts AHR
420 function. Billions of herring in the North Atlantic Ocean are homozygous for Thr346
421 (Extended Data Fig. 8b). Genetic adaptation in a poikilotherm organism exposed to a wide
17 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
422 range of water temperatures (Extended Data Fig. 1) must involve mutations affecting protein
423 function in a temperature-dependent manner. Interestingly, residue 346 occurs in the ligand-
424 binding PAS-B domain and Ser346Thr may contribute to cold adaptation by affecting ligand
425 binding as it dominates in populations exposed to the coldest waters (Extended Fig. 8).
426 The importance of inversions underlying phenotypic variation and ecological
427 adaptation is a hot topic in evolutionary genomics40. Inversions suppress recombination
428 facilitating the evolution of ‘supergenes’ composed of a cluster of co-adapted alleles.
429 Characterization of inversions using short read sequence data is challenging because inversion
430 breakpoints are often embedded in repeat regions. Here we have successfully used PacBio
431 long reads and generated conclusive evidence for inversions on Chr6 and 17 in herring (Fig.
432 6), and suggestive evidence for a third one on Chr23. Furthermore, the individual used for
433 PacBio sequencing was homozygous for the reference allele at the inversion on Chr12
434 documented in our recent study8. The characteristic features of these four inversion
435 polymorphisms are that (i) the inversion breakpoints do not disrupt coding sequence
436 (conclusive evidence only for Chr6 and 17 so far), (ii) no allele is lethal in the homozygous
437 condition, (iii) the inversions suppress recombination but not completely, as we find evidence
438 for recombinant haplotypes (Extended Data Fig. 5), and (iv) none of the inversions is of
439 recent origin as indicated by substantial nucleotide diversity within each haplotype group, in
440 the range 0.11-0.31% (Extended Data Fig. 7). The strong genetic differentiation among
441 populations for each of the four inversions (Extended Data Fig. 6) implies that these are
442 supergenes important for environmental adaptation. It is tempting to speculate that several of
443 these, if not all, are directly or indirectly related to the water temperature at spawning due to
444 the relatively strong correlation between water temperatures and haplotype frequencies
445 (Extended Data Figs 1 and 6). The higher nucleotide diversity for the Southern haplotype
446 groups at each locus suggests that these may represent the ancestral state and that the derived
18 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
447 alleles provided a selective advantage during an expansion of Atlantic herring into colder
448 waters.
449 In conclusion, ecological adaptation in the Atlantic herring is in conflict with the
450 infinitesimal model due to large shifts in allele frequencies at loci under selection. The
451 underlying cause is both (i) sequence variants of major importance, like Phe261Tyr in
452 rhodopsin 36, and (ii) haplotypes of major importance carrying multiple sequence changes.
453 The most extreme form of the latter is megabase inversions/supergenes such as those reported
454 here (Figs 4 and 6).
455
456 Methods
457 Sample collection and genome resequencing. The study is based on sequence data
458 generated in this study as well as those reported in our previous studies (Tables S1 and S2).
459 Tissue from 35-100 fish were collected from different localities in the Baltic Sea, Skagerrak,
460 Kattegat, North Sea and the waters surrounding Ireland and Britain, the Atlantic Ocean, and
461 the Pacific Ocean. Genomic DNA was prepared using standard methods, pooled and
462 individual sequencing was carried out as previously described3. Illumina 2 x150bp short read
463 sequences for the new populations in this study were analysed together with read sets that
464 were generated for populations used previously.
465 Alignment and variant calling. Paired-end reads including the data from previous studies3,7
466 were aligned to the reference genome8 using BWA v0.7.17 component BWA-MEM41.
467 Postprocessing of alignment was handled with Samtools v1.642. Merging libraries of the same
468 sample and marking duplicates were done using Picard toolkit v2.10.343. Raw genomic
469 variants were called using GATK UnifiedGenotyper for pooled sequencing data and GATK
470 HaplotypeCaller for individual sequencing data44. Stringent filtration of the raw variants was
471 done following the criteria “QD < 5.0 || FS > 40.0 || MQ < 50.0 || MQRankSum < -4.0 ||
19 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
472 ReadPosRankSum < -3.0 || ReadPosRankSum > 3.0 || DP < 100 || DP > 5400” for pooled data
473 and “QUAL < 100 || MQ < 50.0 || MQRankSum < -4.0 || ReadPosRankSum < -4.0 ||
474 ReadPosRankSum > 4.0 || QD < 5.0 || FS > 40.0 || DP < 100 || DP > 5500” for individual data
475 using GATK VariantFiltration and SelectVariants. All the cutoffs were determined according
476 to the genome-wide distribution of each parameter. To explore haplotypes in the downstream
477 studies, we phased the individual data using BEAGLE v4.045.
478 Analysis of population divergence. The allele frequency of pool-seq data are calculated
479 based on the counts of reference and alternate reads at each polymorphic site. However, using
480 raw read count could introduce bias in particular population or genomic regions, due to varied
481 sample size across the pools or varied coverage along the genome. To minimize the technical
482 bias, we carried out the Neff allele count correction46,47 before the calculation of allele
483 frequency, following the formula: Neff = (2n * RD- 1) / (2n + RD), where RD is raw read
484 depth and n is the number of individuals in a given pool.
485 To examine population structure and genetic differentiation among the pools, we
486 performed Principal Component Analysis (PCA) based on corrected allele frequencies of two
487 sets of markers showing (i) no significant genetic differentiation and (ii) highly significant
488 genetic differentiation. The two sets of markers were established based on the genome-wide
489 distribution of the standard deviation (std) of allele frequencies across populations. Based on
490 the genome-wide distribution, we defined SNPs with 0.02 ≤ std ≤ 0.08 and std ≥ 0.2 as
491 undifferentiated markers and highly differentiated markers, respectively (Fig. S2). Minor
492 allele frequency was required to be > 0.01 in order to exclude monomorphic sites and rare
493 variants. We further down-sampled the datasets by retaining only one marker within every 1
494 kb of the genome for low differentiated markers and 10 kb for highly differentiated markers,
495 in order to weaken the weight from genetically linked regions; the difference for the two sets
496 are justified because linkage disequilibrium is decaying rapidly in the herring genome except
20 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
497 in some of the regions under selection3. In the final PCA, we used 169,394 SNPs as
498 undifferentiated markers and 794 as highly differentiated markers. Sample 38 was excluded
499 from the PCA analysis due to a potential sequencing quality issue.
500 Additionally, we constructed a genetic distance tree based on ~3.2 million whole-
501 genome SNPs after filtering out missing data, using PHYLIP package v3.69648. We randomly
502 sampled 1,000 replicate datasets using Seqboot bootstrap method and computed pairwise Nei
503 distances using Genedist. For each replicate, a neighbor-joining tree was inferred using
504 Neighbor. A final consensus tree was produced from all trees with Consensus using the
505 Majority Rule method. Tree visualization was done in FigTree v1.4.449.
506 Pairwise FST in 5 kb sliding windows was calculated using samtools42 and
507 PoPoolation250 with the following pipeline and parameters: samtools mpileup -B -q 20 -b;
508 java -ea -Xmx40g -jar mpileup2sync.jar --fastq-type sanger --min-qual 20 --threads 5; perl
509 fst-sliding.pl --min-count 10 --min-coverage 10 --max-coverage 500 --min-covered-fraction
510 0.7 --window-size 5000 --step-size 5000. The sample size of each pool was provided to
511 correct the sequencing depth.
512 Identifying genetic signature of selection. To maximize the statistical power and minimize
513 the within-population variance, we grouped populations into superpools according to their
514 environmental variables or phenotype, and made contrasts between superpools (Table S3).
515 For each contrast, we summed up the read counts separately for reference and alternate alleles
516 in each superpool, and performed 2 2 contingency χ2 test between superpools for every
517 SNP. Bonferroni correction of the P values was calculated in each contrast, and P < 1 10-10
518 was determined as the significance threshold. We clustered the signals as independent loci as
519 described3, by requiring that such regions (1) should be at least 100 bp in length, (2) have at
520 least two SNPs in each contrast reaching a significance threshold of P < 1 10-20 in the χ2 test,
521 and (3) have a minimal distance of 100 kb from the neighboring loci. We summarized the
21 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
522 independent loci in each contrast and intersected the loci between the two replicate contrasts
523 of adaptation to salinity and spawning season.
524 Population genetic analysis of putative structural variants. Diagnostic markers at the four
525 putative structural variants on Chr6, 12, 17 and 23 were selected for investigating allele
526 haplotype frequencies across populations in pooled and individual sequencing data.
527 Independent Neighbor-joining trees were constructed based on all SNPs from each of the
528 putative structural variants using PLINK v1.90b4.951 and neighbor from the PHYLIP package
529 v3.69648. We estimated nucleotide diversity of each haplotype at the SVs using vcftools
530 v0.1.1652.
531 PacBio long read sequencing for characterization of inversion breakpoints. A male
532 Atlantic herring was captured in November 25, 2019 in the Celtic Sea (N51.6, W6.5) and
533 flash frozen in liquid nitrogen. High molecular weight DNA was extracted from testis using
534 the Circulomics Nanobind Tissue Big DNA Kit (NB-900-701-001). Size selection for
535 molecules over 40 kb using BluePippin (SageScience), and libraries for PacBio sequencing
536 was made using the manufacturer’s protocol. Sequencing was carried out on one PacBio
537 Sequel II 8M SMRT Cell for 15 h to generate 125 Gb CLR sequence data. Subreads were
538 aligned to the herring reference assembly8 using Minimap2 aligner53. Only reads with
539 mapping quality greater than 20 were used for further analysis. Average genome coverage
540 was measured using Samtools42 and was found to be 110x. Structural variants were called
541 with Sniffles54. The alignment was viewed using IGV55 and Ribbon56 for the inversion signal.
542 Mummer was used to build the dotplot57.
543 Data availability statement. The sequence data generated in this study is available in
544 Bioproject PRJNA642736.
22 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
545 Code availability statement. The analyses of data have been carried out with publicly
546 available software and all are cited in the Methods section. Custom scripts used are available
547 in Github (https://github.com/Fan-Han/Population-analysis-with-pooled-data)
548
549 Acknowledgements. We are grateful to Erik Enbody for insightful comments on the
550 manuscript and the WESTHER project, the Irish Marine Institute, Marine Scotland Science,
551 the Northern Ireland Agri-Food and Biosciences Institute, H. Ojaveer and Thomas Gröhsler
552 (Thünen Institute of Baltic Sea Fisheries, Rostock, Germany) for sample collections. The
553 study was supported by the Knut and Alice Wallenberg Foundation (to LA), Vetenskapsrådet
554 (to LA), Research Council of Norway project 254774, GENSINC (to AF, FB and LA)
555 AquaCRISPR (ANR-16-COFA-0004-01) (to AH) , and Danish Ministry for Food,
556 Agriculture and Fisheries and the European Fisheries Fund: contract no. 33113‐B‐19‐154)
557 (to DB). The National Genomics Infrastructure (NGI)/Uppsala Genome Center and
558 UPPMAX provided service in massive parallel sequencing and computational infrastructure.
559 Work performed at NGI/Uppsala Genome Center has been funded by RFI/VR and Science for
560 Life Laboratory, Sweden.
561 Author contributions. LA and AF conceived the study. FH was responsible for
562 bioinformatic analysis of population data. MJ was responsible for analysis of PacBio data.
563 MEP, LS, AF-P, and BD contributed to the bioinformatic analysis. DB, FB, MC, GD, EF, and
564 AF contributed with samples and to the interpretation of population data. LA and FH wrote
565 the paper with contributions from all other authors. All authors approved the manuscript
566 before submission.
567 Competing interest statement. The authors declare no competing interest.
23 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
568
569 References
570 1. Food and Agriculture Organization of the United Nations Statistics Division (FAO).
571 FAO Yearbook: Fishery and Aquaculture Statistics of 2015. FAO Annual Yearbook
572 (2017). doi:10.5860/CHOICE.50-5350
573 2. Feng, C. et al. Moderate nucleotide diversity in the Atlantic herring is associated with a
574 low mutation rate. Elife 6, e23907 (2017).
575 3. Barrio, A. M. et al. The genetic basis for ecological adaptation of the Atlantic herring
576 revealed by genome sequencing. Elife 5, e12081 (2016).
577 4. Lamichhaney, S. et al. Population-scale sequencing reveals genetic differentiation due
578 to local adaptation in Atlantic herring. Proc Natl Acad Sci U S A 109, 19345–19350
579 (2012).
580 5. Limborg, M. T. et al. Environmental selection on transcriptome-derived SNPs in a high
581 gene flow marine fish, the Atlantic herring (Clupea harengus). Mol Ecol 21, 3686–
582 3703 (2012).
583 6. Fisher, R. A. The Correlation between Relatives on the Supposition of Mendelian
584 Inheritance. Trans R Soc Edinburgh 52, 399–433 (1918).
585 7. Lamichhaney, S. et al. Parallel adaptive evolution of geographically distant herring
586 populations on both sides of the North Atlantic Ocean. Proc Natl Acad Sci U S A 114,
587 E3452–E3461 (2017).
588 8. Pettersson, M. E. et al. A chromosome-level assembly of the Atlantic herring genome-
589 detection of a supergene and other signals of selection. Genome Res 29, 1919–1928
590 (2019).
591 9. Eggers, F. et al. Seasonal dynamics of atlantic herring (Clupea harengus L.)
592 populations spawning in the vicinity of marginal habitats. PLoS One 9, (2014).
24 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
593 10. Clausen, L. A. W., Bekkevold, D., Hatfield, E. M. C. & Mosegaard, H. Application and
594 validation of otolith microstructure as a stock identification method in mixed Atlantic
595 herring (Clupea harengus) stocks in the North Sea and western Baltic. ICES J Mar Sci
596 64, 377–385 (2007).
597 11. McQuinn, I. H. Metapopulations and the Atlantic herring. Rev Fish Biol Fish 7, 297–
598 329 (1997).
599 12. Bekkevold, D. et al. Divergent origins of sympatric herring population components
600 determined using genetic mixture analysis. Mar Ecol Prog Ser 337, 187–196 (2007).
601 13. Berg, F., Østgaard, H. D., Slotte, A., Andersson, L. & Folkvord, A. A combination of
602 genetic and phenotypic characterization of spring- and autumn-spawning herring
603 suggests gene flow between populations. ICES J Mar Sci (2020).
604 doi:10.1093/icesjms/fsaa046
605 14. Manzon, L. A. The role of prolactin in fish osmoregulation: A review. General and
606 Comparative Endocrinology 125, 291–310 (2002).
607 15. Nakane, Y. et al. The saccus vasculosus of fish is a sensor of seasonal changes in day
608 length. Nat Commun 4, (2013).
609 16. Nakane, Y. & Yoshimura, T. Photoperiodic Regulation of Reproduction in Vertebrates.
610 Annu Rev Anim Biosci 7, 173–194 (2019).
611 17. Melamed, P. et al. Gonadotrophin-Releasing Hormone Signalling Downstream of
612 Calmodulin. Journal of Neuroendocrinology 24, 1463–1475 (2012).
613 18. Santos, E. M. et al. Molecular basis of sex and reproductive status in breeding
614 zebrafish. Physiol Genomics 30, 111–122 (2007).
615 19. Maddox, D. M. et al. A mutation in Syne2 causes early retinal defects in
616 photoreceptors, secondary neurons, and Müller Glia. Investig Ophthalmol Vis Sci 56,
617 3776–3787 (2015).
25 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
618 20. Johnston, I. A., Vieira, V. L. A. & Temple, G. K. Functional consequences population
619 differences in the developmental plasticity of muscle to temperature in Atlantic herring
620 Clupea harengus. Mar Ecol Prog Ser 213, 285–300 (2001).
621 21. ICES. Herring (Clupea harengus) in subareas 1, 2, and 5, and in divisions 4.a and
622 14.a, Norwegian spring-spawning herring (the Northeast Atlantic and the Arctic
623 Ocean). ICES Advice on fishing opportunities, catch, and effort Northeast Atlantic and
624 Arctic Ocean ecoregions her.27.1-2, (2019).
625 22. Dragesund, O., Johannessen, A. & Ulltang, Ø. Varlation in migration and abundance of
626 norwegian spring spawning herring (clupea harengus l.). Sarsia 82, 97–105 (1997).
627 23. Ferreira-Martins, D., Coimbra, J., Antunes, C. & Wilson, J. M. Effects of salinity on
628 upstream-migrating, spawning sea lamprey, Petromyzon marinus. Conserv Physiol 4,
629 (2016).
630 24. Ferrara, A. M. et al. Homozygous thyroid hormone receptor β-gene mutations in
631 resistance to thyroid hormone: Three new cases and review of the literature. J Clin
632 Endocrinol Metab 97, 1328–1336 (2012).
633 25. Deveau, C. et al. Thyroid hormone receptor beta mutations alter photoreceptor
634 development and function in Danio rerio (zebrafish). bioRxiv 1–31 (2019).
635 doi:doi.org/10.1101/800821
636 26. Min, Z. et al. Asymmetrical methyltransferase PRMT3 regulates human mesenchymal
637 stem cell osteogenesis via miR-3648. Cell Death Dis 10, (2019).
638 27. Andersson, L., Ryman, N., Rosenberg, R. & Ståhl, G. Genetic variability in Atlantic
639 herring (Clupea harengus harengus): description of protein loci and population data.
640 Hereditas 95, 69–78 (1981).
641 28. Ryman, N., Lagercrantz, U., Andersson, L., Chakraborty, R. & Andersson, L. Lack of
642 correspondence between genetic and morphologic variability patterns in atlantic
26 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
643 herring (Clupea harengus). Heredity (Edinb) 53, 687–704 (1984).
644 29. Larsson, L. C., Laikre, L., André, C., Dahlgren, T. G. & Ryman, N. Temporally stable
645 genetic structure of heavily exploited Atlantic herring (Clupea harengus) in Swedish
646 waters. Heredity (Edinb) 104, 40–51 (2010).
647 30. Cope, J. M. & Punt, A. E. Reconciling stock assessment and management scales under
648 conditions of spatially varying catch histories. Fish Res 107, 22–38 (2011).
649 31. Hintzen, N. T. et al. Managing a complex population structure: Exploring the
650 importance of information from fisheries-independent sources. ICES J Mar Sci 72,
651 528–542 (2014).
652 32. Goddard, M. E., Kemper, K. E., MacLeod, I. M., Chamberlain, A. J. & Hayes, B. J.
653 Genetics of complex traits: Prediction of phenotype, identification of causal
654 polymorphisms and genetic architecture. Proc R Soc B Biol Sci 283, (2016).
655 33. Wood, A. R. et al. Defining the role of common variation in the genomic and
656 biological architecture of adult human height. Nat Genet 46, 1173–1186 (2014).
657 34. Bulmer, M. G. The Effect of Selection on Genetic Variability. Am Nat (1971).
658 doi:10.1086/282718
659 35. Andrén, T. et al. The Development of the Baltic Sea Basin During the Last 130 ka. in
660 75–97 (2011). doi:10.1007/978-3-642-17220-5_4
661 36. Hill, J. et al. Recurrent convergent evolution at amino acid residue 261 in fish
662 rhodopsin. Proc Natl Acad Sci U S A 116, 18473–18478 (2019).
663 37. Schulte, K. W., Green, E., Wilz, A., Platten, M. & Daumke, O. Structural Basis for
664 Aryl Hydrocarbon Receptor-Mediated Gene Activation. Structure 25, 1025-1033.e3
665 (2017).
666 38. Tischkau, S. A. Mechanisms of circadian clock interactions with aryl hydrocarbon
667 receptor signalling. Eur J Neurosci 51, 379–395 (2020).
27 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
668 39. Oziolor, E. M. et al. Adaptive introgression enables evolutionary rescue from extreme
669 environmental pollution. Science 364, 455–457 (2019).
670 40. Wellenreuther, M. & Bernatchez, L. Eco-Evolutionary Genomics of Chromosomal
671 Inversions. Trends in Ecology and Evolution 33, 427–440 (2018).
672 41. Li, H. Aligning sequence reads, clone sequences and assembly contigs with BWA-
673 MEM. arXiv 00, 1–3 (2013).
674 42. Li, H. et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics 25,
675 2078–2079 (2009).
676 43. Github. Picard toolkit. Broad Institute, GitHub Repos
677 http://broadinstitute.github.io/picard/ (2019).
678 44. McKenna, A. et al. The genome analysis toolkit: A MapReduce framework for
679 analyzing next-generation DNA sequencing data. Genome Res 20, 1297–1303 (2010).
680 45. Browning, S. R. & Browning, B. L. Rapid and accurate haplotype phasing and
681 missing-data inference for whole-genome association studies by use of localized
682 haplotype clustering. Am J Hum Genet 81, 1084–1097 (2007).
683 46. Bergland, A. O., Behrman, E. L., O’Brien, K. R., Schmidt, P. S. & Petrov, D. A.
684 Genomic Evidence of Rapid and Stable Adaptive Oscillations over Seasonal Time
685 Scales in Drosophila. PLoS Genet 10, e1004775 (2014).
686 47. Fuentes-Pardo, A. P. et al. Adaptation to seasonal reproduction and thermal minima-
687 related factors drives fine-scale divergence despite gene flow in Atlantic herring
688 populations. bioRxiv 1, 1–48 (2019).
689 48. Felsenstein, J. PHYLIP (Phylogeny inference package), 3.572 (PowerPC). Cladistics 5,
690 164–166 (1995).
691 49. Rambaut, A. FigTree, a graphical viewer of phylogenetic trees (tree.bio.ed.ac.
692 uk/software/figtree). Inst Evol Biol Univ Edinburgh 1.4.2, (2014).
28 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
693 50. Kofler, R., Pandey, R. V. & Schlötterer, C. PoPoolation2: Identifying differentiation
694 between populations using sequencing of pooled DNA samples (Pool-Seq).
695 Bioinformatics 27, 3435–3436 (2011).
696 51. Purcell, S. et al. PLINK: A tool set for whole-genome association and population-
697 based linkage analyses. Am J Hum Genet 81, 559–575 (2007).
698 52. Danecek, P. et al. The variant call format and VCFtools. Bioinformatics 27, 2156–2158
699 (2011).
700 53. Li, H. Minimap2: Pairwise alignment for nucleotide sequences. Bioinformatics 34,
701 3094–3100 (2018).
702 54. Sedlazeck, F. J. et al. Accurate detection of complex structural variations using single-
703 molecule sequencing. Nat Methods 15, 461–468 (2018).
704 55. Robinson, J. T. et al. Integrative genomics viewer. Nature Biotechnology 29, 24–26
705 (2011).
706 56. Nattestad, M., Chin, C.-S. & Schatz, M. C. Ribbon: Visualizing complex genome
707 alignments and structural variation. bioRxiv 0344, 82123 (2016).
708 57. Delcher, A. L. Fast algorithms for large-scale genome alignment and comparison.
709 Nucleic Acids Res 30, 2478–2483 (2002).
710
711 Figure legends
712 Fig. 1. Geographic distribution of sampled pools of Atlantic herring. Color code of each
713 population refers to the spawning season that a given pool was sampled.
714 Fig. 2. Principal component analysis (PCA) of herring populations. (a) PCA based on 22,195
715 genetically undifferentiated markers. (b) PCA based on 794 markers that are likely under
716 natural selection. Color of each population represents spawning season, as coded in Figure 1.
717 Inset bar plot indicates the percentage of explained variance by each principal component.
29 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
718 Fig. 3. Genetic signals associated with ecological adaptation. (a) Genetic differentiation
719 between superpools of Baltic and Atlantic herring in spring spawners (above) and in autumn
720 spawners (below) and (b) a zoomed-in profile for Chr12. (c) Genetic differentiation between
721 super-pools of spring and autumn herring in Baltic (above) and in Atlantic herring (below)
722 and (d) a zoomed-in profile on Chr15. Red dots indicate the signals sharing in both contrasts.
723 * Peak is located in intergenic region. The closest gene is labeled.
724 ** An identified locus that is associated with adaptation to the red-shifted light environment
725 in the Baltic Sea.
726 Fig. 4. Genetic differentiation between herring populations from Ireland and Britain vs. other
727 populations from the Northeast Atlantic. (a) Putative structural variants on Chr6, 12, 17 and
728 23 are indicated with red stars. (b) to (e) are zoomed-in profiles on the corresponding
729 chromosomes. Strong genetic differentiation is observed in the following intervals: Chr6: 22.2
730 – 24.8 Mb, Chr12: 17.8 – 25.6 Mb, Chr17: 25.8 – 27.5 Mb, and Chr23: 16.3 – 17.5 Mb.
731 Fig. 5. Genetic differentiation between Atlantic herring from the North Atlantic vs. waters
732 around Ireland, British, and the transition zone. (a) Genome-wide screen. (b) Genetic
733 differentiation on Chr2 and a zoom-in profile of AHR locus. Red dots refer to the four
734 missense mutations.
735 Fig. 6. PacBio long reads defining inversion breakpoints on chromosomes 6 and 17. (a)
736 Chromosome 6. The orientation of an informative read is indicated with arrows above the
737 Chr6 reference panel. The breakpoints are indicated with dotted blue lines. The region just
738 outside the proximal breakpoint is misassembled in the reference and results in a 1.7 kb
739 pseudo-inversion (marked by an asterisk below the read panel) flanked by 499 bp of Ns. It
740 also results in the finding that nearly 10 kb of the PacBio read maps to unplaced_scaffold699
741 (shown in yellow colored box) and nearly 5 kb (indicated with hash symbol below the read
742 panel) that is not present in the assembly. Genes surrounding the breakpoints are indicated
30 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
743 above the reference panel. CCDC102A: coiled-coil domain containing 102A, DOK4: docking
744 protein 4, PRMT3: protein arginine methyltransferase 3, CPNE7: copine-7-like, BAX:
745 apoptosis regulator BAX-like, FGF19: fibroblast growth factor 19. (b) Chromosome 17. The
746 correspondence between the PacBio reads and the reference assembly is indicated with arrows
747 above the reference panel. The breakpoints are indicated with dotted blue lines. The overlap
748 of the read to both the proximal and distal breakpoints is shaded in dark green and is caused
749 by a 5 kb duplication that is present before the proximal and distal breakpoints. The dotplot
750 illustrating this duplication is shown in the inserted graph. Genes in the near vicinity of the
751 breakpoints are shown above the reference sequence. SUGCT: succinyl –CoA: glutarate-CoA
752 transferase, UBA5: ubiquitin-like modifier activating enzyme 5, NPHP3: nephronophthisis 3,
753 ZMYND11: zinc finger, MYND-type containing 11, LARP4B: La ribonucleoprotein 4B
754 homolog, DIP2CA: disco-interacting protein 2 homolog Ca, VIPR2: vasoactive intestinal
755 peptide like receptor 2.
756
757 Extended Data Fig. 1. Variation in sea water temperatures among seasons. Seasonal sea
758 water temperature profiles across the North Atlantic and the Baltic Sea. Herring sampling
759 locations are overlaid on top of the temperature data. Climatic data layers at a 0.25° spatial
760 resolution were retrieved from the World Ocean Atlas 2018 (WOA18)
761 ((https://www.nodc.noaa.gov/cgi-bin/OC5/woa18/woa18.pl?parameter=t) oceanographic
762 database created and maintained by the National Oceanic and Atmospheric Administration
763 (NOAA). Data layers correspond to average temperature values from 2005 to 2017
764 aggregated in four seasons as follows: winter (January to March), spring (April to June),
765 summer (July to September), and autumn (October to December). Map plotting was
766 performed using the R package (https://www.r-project.org). Colors represent degree Celsius.
31 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
767 Extended Data Fig. 2. Phylogenetic reconstruction. (a) Neighbor-joining tree of 53 Atlantic
768 herring populations based on allele frequency of genome-wide SNPs. (b) Geographical
769 distribution of populations in relation to phylogenetic clustering. Color code of each
770 population refers to the cluster grouped in NJ tree.
771 Extended Data Fig. 3. Genetic differentiation around (a) PRLR gene and (b) SOX11B gene
772 in each contrast.
773 Extended Data Fig. 4. Allele frequency of diagnostic SNPs at four putative structural
774 variants across populations of Atlantic herring and one Pacific outgroup.
775 Extended Data Fig. 5. Heatmap of the genotypes for diagnostic SNPs based on individual
776 whole genome sequencing data at four putative structural variants.
777 Extended Data Fig. 6. Frequencies of two estimated haplotypes in each population at four
778 putative structural variants on Chr6 (a), Chr12 (b), Chr17 (c), and Chr23 (d) across the
779 geographical distribution.
780 Extended Data Fig. 7. Neighbor-joining tree based on genotypes of all SNPs in each of the
781 four putative structural variants respectively on Chr6 (a), Chr12 (b), Chr17 (c) and Chr23 (d).
782 Average nucleotide diversity was estimated for S and N haplotypes in each tree.
783 Extended Data Fig. 8. The missense mutation (Ser346Thr) and multi-species alignment at
784 AHR locus. (a) Allele frequency of the missense mutation (Ser346Thr) in different
785 populations. (b) Amino acid alignment of AHR protein across multiple species.
786 Extended Data Fig. 9: Heatmap of SNP frequencies for all pooled samples. The SNPs
787 included are those with absolute delta allele frequency above 0.4 in the interval 6.35 Mb to
788 6.37 Mb on chromosome 19. The color side-bar indicates the frequency of the reference
789 allele, and the top bar highlights (in purple) the one non-synonymous SNP in the set:
790 chr19_6359267. This polymorphism causes a missense mutation Gln40His in THRB gene
791 (ENSCHAG00000003500). White cells represent missing data.
32 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
792 Supplementary Fig. 1. Pairwise FST among 53 populations of Atlantic herring.
793 Supplementary Fig. 2. Histogram of genome-wide standard deviation of allele frequencies
794 across 52 Atlantic and Baltic populations. We selected 22,195 undifferentiated markers and
795 794 markers that are likely under natural selection based on the cutoffs (SD < 0.02 and SD >
796 0.2).
797 Supplementary Table 1. Specimen information of 53 pooled samples.
798 Supplementary Table 2. Specimen information of 55 individual sequencing samples.
799 Supplementary Table 3. Definition of how superpools were formed by pooling population
800 samples of major groups of herring. Sample names correspond to those given in
801 Supplementary Table 1.
802 Supplementary Table 4. Genes within inverted regions and 200 kb flanking inversion
803 breakpoints on chromosome 6 an 17 in Atlantic herring.
33 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer● review)Autumn is the● author/funder.Spring● All rights Summer reserved.● No Winter reuse allowed● withoutMixed permission.
70 N
● 41
● 45 ● 65 N 1 44 ●
43 9 ● ● 32 6 ● 11 ● 31 ● ●● 60 N 51 30 10 8 37 52 ● 28 ● 2 46 ● 24 12 7 3 ● 53 ● ●● ● 26 ●● ● 25 ● ● 42 27 ● ● 13 4 5 29 19 ● ● 14 33 22 ● 23 ●● 17 55 N 35 ● 20 ● ● ● 18 Latitude (°) ● 34 North Atlantic Ocean 21 16 15 ● 47 ● 50 N 36 48 ● ● 39 40 ● ● 38 ● 45 N 50 ● 49
60 W 40 W 20 W 0 20 E Longitude (°)
Figure 1. bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint a (which was not certified by peer review) is the author/funder. All brights reserved. No reuse allowed without permission.
7 8 1 10 ● ●● 40 ● ●● Baltic Sea ●37 ●● 14 Atlantic Ocean 30 12 ● (autumn) 17 6 11 53 50 20 ● Atlantic Ocean ● ● ● 45 10 ● ● (autumn) 13 ● ● ● 9 4 22 39 49 37 0 ● ● ● ● ● 48 2 ● ● Baltic Sea ●● ● ● ● ● ● 40 ●● 1 2 3 4 5 6 7 8 9 10 (spring) 3 5 ● ●●● 52 35 50 49 19 44 51●● ●●●●●● 42 47 48 34 33 45 30 Baltic Sea ● ● 41 ● 41 36 Baltic Sea 51 ● 46 (spring) 44● ● 8 (autumn) 39 27 28 1 11 ●● 40 ● 14 ● 26 ● ● Atlantic Ocean ● ● 4 52 ● ●● 6 22 ● 31 ●● ● ● (spring) 18 32 ● ● PC2 ● PC2 10 5 ● ● Norwegian ● 9 Ireland and 2 19 ● 23 fjords 43 17 ●● Britain 24 ● 16 ● ● 3 Norwegian ● 7 46 20 fjords 53 ● 16 ● ● 29 ● 30 ● 32 ● ● 13 12 ● 26 ● ● 21 25 18 ● 43 31 ● ● ● 6 ● 20 42 ● 15 ●● 27 28 Transition zone ● ● ● 4 ● ● ● 23 33 35 ● ● 29 ● ● ● 21 ● 2 ● ● 24 Transition zone 34 ● ● ● 0 ● 15 ● 47 1 2 3 4 5 6 7 8 9 10 25 36 PC1 PC1
Figure 2. a bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this versionb posted July 15, 2020. The copyright holder for this preprint Baltic(which vs. was Atlantic not certified superpools by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 140 LRRC8C NDUFAF2 SCD5 PRLR ENOPH1 RHO** LOC105899468 CBLN3 75 PRLR Spring 70 50 PLPP1A
AFF1* 25 Spring - log10 (P) 1 3 5 7 9 11 13 15 17 19 21 23 25 0 2 4 6 8 10 12 14 16 18 20 22 24 26 Chr 12 10 20 (Mb) -log10 (P)
25 Autumn
Autumn 70 LOC105892234 50 c spring vs. autumn superpools d TSHR TSHR MYHC 100 SOX11B* CEP128 ALLC GTF2A1
90 ESR2A SYNE2 HERPUD2 Baltic Sea NSD2* 60 50 CALM1B Baltic Sea 30
0 1 3 5 7 9 11 13 15 17 19 21 23 25 Chr 15 8 9 10 11 (Mb) 2 4 6 8 10 12 14 16 18 20 22 24 26 -log10 (P) -log10 (P)
30 40 Atlantic Ocean
Atlantic Ocean 60
90 80
Figure 3. a bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this versionb posted July 15, 2020. The copyright cholder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 7.8 Mb 80 80
80 40 2.7 Mb 40 - log10 (P) - log10 (P) 20 20 40 TSHR ESR2A CD209 0 0 SLC12A2 SYNE2 5 10 15 20 25 30 (Mb) 5 10 15 20 25 30 (Mb) HERPUD2 d Chr 6 e Chr 12 AHR NCL1 - log10 (P) 80 80 20
40 40 2.2 Mb 1.2 Mb 0 - log10 (P) 1 3 5 7 9 11 13 15 17 19 21 23 25 20 - log10 (P) 20 2 4 6 8 10 12 14 16 18 20 22 24 26
0 0 5 10 15 20 25 (Mb) 5 10 15 20 25 (Mb) Chr 17 Chr 23 Figure 4. bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint a (which was not certified by peer review) is the author/funder. All rightsb reserved. No reuse allowed without permission. 100 AHR 100 75 AHR 50 25 75 - log10 (P) 0 CILP1 CLEC ESR2A 10 20 30 (Mb) MXG CCNB1 50 SLC12A2 THRB
- log10 (P) 100 Ser346Thr Gln808Pro 25 75 50 Ala299Gly Leu813Met 25 0 - log10 (P) 1 3 5 7 9 11 13 15 17 19 21 23 25 0 2 4 6 8 10 12 14 16 18 20 22 24 26 15805000 15810000 15815000 15820000 (bp) Chr 2 Figure 5. bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint a (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. DOK4 BAX FGF19 A201CDCC PRMT3 CPNE7
pb 630,082,22 pb 0 pb 630,082,22 pb 185,868,42 pB 455,164,13 pb Chr6
PacBio Read un_sc699 # Ns * Ns 0 kb 10 kb 20 kb 30 kb
UBA5 b SUGCT LARP4B ZMYND11 NPHP3 DIP2CA VIPR2
0 bp 25,805,445 bp 27,568,511 bp Chr17 PacBio Read 0 kb 5 kb 10 kb 15 kb 20 kb 4938 distal en d (bp) e distal th Within inversion at inversion Within 0 0 4686 Upstream of inversion proximal end (bp)
Figure 6 WinterbioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint 70 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 41 66 1 45 44 43 9 62 37 31 32 10 6 11 8 52 51 30 28 26 7 2 24 12 5 58 46 42 53 25 27 19 13 3 14 4 29 23 20 22 54 35 33 34 17 21 16 15 18 47
Latitude ( ) 36 50 39 48 46 40 38 50 49 42 0 10 20 30 Longitude ( ) Spring 70
41 66 45 44 1 43 9 62 32 10 37 31 6 11 8 52 51 28 2 24 30 26 12 7 5 58 46 42 53 25 27 19 13 3 14 4 29 23 20 22 33 17 54 35 34 21 16 15 18 36
Latitude ( ) 47 50 48 Temperature (C°) 39 22 46 40 38 50 20 49 18 42 16 0 10 20 30 14 Longitude ( ) 12 Summer 10 70 8 6 41 4 45 66 1 44 2 43 9 6 0 62 32 31 10 37 11 8 52 51 30 28 26 2 24 12 7 5 58 46 42 53 25 27 19 3 14 13 4 29 22 33 34 23 20 17 54 35 18 21 16 15 47 36 Latitude ( ) 50 39 48 46 40 38 50 49 42 0 10 20 30 Longitude ( ) Autumn 70 41 66 45 44 1
43 9 6 32 62 31 10 37 8 11 52 51 30 28 26 2 24 12 7 5 58 46 53 25 27 19 3 42 14 13 4 29 23 20 22 35 33 17 54 34 16 21 15 18 36
Latitude ( ) 47 50 39 48 46 40 38 50 49 42 0 10 20 30 Longitude ( )
Extended Data Fig 1. bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint a (which was not certified● byAutumn peer review)● isSpring the author/funder.● Summer All rights ●reserved. Winter No reuse● allowedMixed without permission. Baltic Sea (spring) 0.002
8
1 17 10 Transition zone 26 14 9 7 12 37 15 6
18 3
16 2 13 21 20 25 48 Atlantic Ocean 49 50 (autumn) 24 38 23 39 27 40 30 41 28 44 45 4 5 22 Atlantic Ocean 29 11 19 (spring) 43 Baltic Sea 32 31 (autumn) Norwegian 42 fjords 33
52 46 35
36 53 34
51 47 Ireland and Britain
b
70 N
41 ●
● ● 1 65 N 44 45 ● 43 9 ● 32 31 ● 6 ● 11 ● ● ●●● 60 N 37 51 30 10 ● 8 52 ● 28 ● 46 ● 24 12 7 ● 3 ● 53 ● ●● ● ●●● 2 ● 25 ● 26 ●● 29 ● 13 4 5 42 27 19 14 33 ● ● ● 23 22 ●● 17 55 N ● 20 ● ● 35 ● ● 18 Latitude (°) 34 21 16 15 ● 36 47 ● 50 N 48 39 ● ● 40 ● ● 38 ● 45 N 50 49 ●
60 W 40 W 20 W 0 20 E Longitude (°)
Extended Data Fig 2. bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint a (which was not certified by peer review)Baltic is the author/funder. vs. Atlantic All rights superpools reserved. No reuse allowed without permission. PRLR
● Spring ● ● ● ●
● ●
● ● ● ●● ● ● ● ● ● ● ● 75 ● ● ● ● ● ● ● ● ● ●
● ● ● ● ● ● ● ● ● ●
●
●
● ● ●
● ●
● ● ●
● ●
50 ● ●
● ● ●
● ●
● ●
● ● ● ● ● ● ●
● ● ● ● ● ● 25 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 0 ●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
Chr12 2.27 2.28 2.29 2.30 (Mb)
● ● ● ● ● ● ●
● ●
● ● ● ●
●
●
●
● ●
●
●
●
● ● ●
● ● ● ● ● ● ●
● ● ● ●
● ●
● ● ● ●
●
●
● ● ●
●
●
● ●
●
● ● ● ● ● ● ● ● ●
● ● ● ●
● ● ●
● ● ● ●
● ●
● ●
● ●
●
●
● ●
●
●
●
● ● ●
● ●
● ●
● ● ● ●
● ● ●
● ● ● ● ●
●
●
● ●
●
● ● ●
●
●
●
●
● ●
●
●
●
●
● ● ●
●
●
● ●
● ● ●
● ● ● ●
● ● ●
●
●
● ●
●
●
●
●
● ● ● ● ● ● ●
●
●
● ●
●
● ●
●
●
0 ●
● ● ● ●
● ●
● ●
●
●
● ●
● ● ●
●
●
●
●
●
●
●
● ●
-log10 (P) ● ●
● ●
●
●
●
●
●
●
●
●
●
●
● ●
●
● ● ●
●
● ●
● ● ●
●
●
●
●
●
●
●
●
●
●
●
● ● ●
●
●
●
●
● ●
●
●
●
● ●
●
●
● ●
● ●
●
●
●
●
●
●
● ●
● ●
●
● ●
●
●
●
●
●
● ●
●
●
●
● ●
●
● ● ●
●
●
●
●
●
● ● ●
●
● ●
●
●
●
●
●
●
● ●
●
● ● ● ●
●
● ●
●
●
●
●
●
25 ●
● ●
●
●
●
● ●
● ●
● ●
●
●
●
●
●
● ●
50
75 Autumn
b spring vs. autumn superpools SOX11B ALLC RPS7 RNASEH1
● Baltic Sea
90
●
●
● ●
●
● 60 ● ●
● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 30 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ●● ● ● ● ●● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●● ●● ● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ●● ●● ● ● ● ● ● ●● ● ● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●● ● ●● ● ● ● ●● ● ● ●● ●● ● ●● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ●● ●● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ●● ● ● ●●● ● ● ● ● ● ● ●●● ● ● ●● ● ● ● ● ● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ●● ●● ●● ● ● ●●● ● ● ● ● ● ● ● ● ● ●●● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ●●● ●●●● ● ●●● ● ●● ● ● ● ● ●●● ● ●● ● ● ● ●● ● ● ●● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ●●● ●● ● ● ● ● ●● ● ● ● ● ● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●● ● ● ● ● ● ● ● ● ● ●● ● ●●●● ● ●● ●● ● ● ● ●● ● ● ● 0 ● ● ● ●● ● ● ● ●● ● ●●●●●●● ● ● ● ● ● ●●●●●●●●●●●● ● ●● ●● ●●● ●● ● ● ● ● ● ● ● ●● ●●● ●● ●● ●●● ● ●●● ●● ● ●● ●● ●● ● ● ●●●● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ●●● ● ● ●● ●● ● ● ● ● ●● ● ●● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ●● ● ●● ● ● ● ●●●●●●●● ● ● ●●● ●● ●● ● ● ●
Chr15 7.75 7.80 7.85 7.90 (Mb)
● ● ● ● ● ● ● ●
● ● ● ● ● ● ● ● ● ●
● ● ● ● ● ●
● ● ●
● ● ● ● ●
●
● ●
● ● ● ● ● ●
● ● ● ● ● ● ● ●
●● ● ●
● ● ● ● ● ●
● ● ● ● ● ● ● ●
● ● ● ●
● ● ● ● ● ● ● ●
● ● ● ● ●
● ● ● ●
●
● ● ● ● ●
● ● ●
● ● ● ●
● ● ● ● ●
● ● ● ● ● ●
● ● ● ● ● ●
● ● ● ● ● ● ● ● ● ●
●● ● ● ● ● ● ● ● ● ●
● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ●
●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●
● ● ● ● ● ● ● ● ● ●
● ● ● ● ● ● ●
● ● ● ● ● ● ● ● ●
● ●
● ● ●
● ● ● ● ● ● ● ●
● ● ● ●
● ●
● ● ● ● ●
● ●
● ● ●
● ● ●
● ● ● ● ● ● ●
● ● ● ● ● ● ● ● ● ●
● ● ● ● ● ●
● ● ● ●
● ● ● ● ● ●
● ● ● ● ● ●
● ● ● ● ● ● ● ●
● ● ● ● ● ●
●
● ● ●
● ●
● ● ● ● ●
● ● ● ●
● ● ● ● ●
● ●
● ● ● ● ●
●
● ●
● ● ●
● ● ●
● ● ● ● ● ● ● ● ●
● ● ● ●
● ● ●
● ● ●
● ● ●
● ● ● ● ● ●
● ● ● ●
● ● ● ●
●
● ● ● ●
● ● ● ●
● ● ● ● ● ● ● ● ● ●
● ● ● ● ●
● ● ● ●
● ●
● ● ● ● ●● ● ● ● ●
● ● ● ● ●
● ● ● ● ● ● ● ●
● ● ● ● ● ● ● ● ●
● ● ● ● ● ● ● ● ● ● ● ● ●
● ● ● ● ●
● ● ● ●
● ● ● ● ●
● ● ● ● ● ● ● ● ●
● ● ● ● ● ● ●
● ● ● ● ●
● ● ●
● ● ● ● ● ●
● ● ● ● ● ●
● ●
●
● ● ●
●
● ● ● ●
● ●
● ● ● ● ● ● ● ● ●
● ● ●
●
●● ● ●
● ●
● ● ● ● ●
● ● ● ●
● ● ●
●
● ● ●
●
● ● ●
●
● ●
●
● ●
● ● ● ● ●
● ●
● ●
●
● ●
● ● ●
● ● ● ● ● ● ● ●
● ● ● ●
● ●
● ● ● ● ●
● ● ● ●
●● ● ● ● ●
● ● ● ●
● ●
●
● ●
● ● ●
● ● ● ● ● ●
● ●
●
● ●
●
● ● ● ● ●
● ● ●
●
● ●
●
● ●
● ●
●
● ● ● ●
● ● ●
● ● ● ●
● ● ● ●
●
●
●
●
●
● ●
●
● ●
● ● ●
● ●
● ●
● ● ●
●
● ● ●
● ● ● ●
● ● ● ● ●
● ● ● ●
●
● ●
● ●
● ●
●
●
● ●
● ● ● ●
● ● ●
● ● ●
● ●
● ●
● ●
● ●
● ● ● ●
●
●
● ● ●
● ● ● ● ●
● ●
●
● ●
● ● ●
●
● ● ●
● ●
● ●
● ●
● ●
● ● ●
●
●
● ● ● ●
●
●
● ● ●
●
●
● ● ● ●
●
● ● ● ● ● ● ● ●
●
●
● ●
●
● ● ●
●
● ●
● ● ● ● ●
● ● ● ●
● ● ●
● ●
● ● ● ●
●
●
●
●
●
● ● ●
● ●
●
● ●
●
● ● ● ● ●
●
●
●
●
● ● ●
● ● ● ● ●
●
● ● ●
●
● ● ● ●
●
● ● ●
●
●
● ● ● ●
● ● ● ● ●
● ●
●
● ● ●
● ●
● ● ●
● ●
●
●
● ●
●
● ● ●
●
●
● ●
●
● ●
● ● ● ●
● ● ● ●
● ●
● ● ●
●● ● ●
● ●
●
●
● ●
● ● ●
● ● ●
● ●
●
●
●
● ●
● ● ● ●
●
●
● ●
● ● ●
●
● ●
● ●
●
●
●
●
● ●
●
●
●
● ● ●
● ● ●
●
● ●
●
●
●
●
●
● ●
●
●
● ● ●
●
0 ●
● ● ●
● ● ● ● ●
●
● ●
● ●
● ●
●
●
●
●
●
●
●
● ●
●
●
●
● ● ●
●
●
●
● ●
● ● ● ● ●
●
●
● ● ● ● ●
● ● ●
●
●
● ● ● ●
●
● ● ●
● ● ●
●
● ● ●
● ●
● ●
●
●
● ● ● ●
● ●
● ●
●
● ●
● ● ●
●
●
● ●
● ●
● ●
● ●
● ● ● ● ●
●
● ●
● ● ●
● ●
● ● ● ●
● ●
●
●
● ●
●
● ●
●
●
● ●
● ● ●
●
●
●
● ●
●
● ●
●
●
● ●
●
●
●
● ● ● ●
●
●
● ●
●
●
●
●
●
● ● ●
●
●
●
●
● ●
●
●
● ●
●
●
●
●
● ● ●
●
● ●
● ●
● ●
● ● ●
● ● ● ●
●
●
●
● ● ●
●
● ● ● ● ●
●
●
● ●
●
●
●
● ● ●
● ●
●
● ●
● ●
● ●
●
●
● ● ● ●
●
●
●
-log10 (P) ●
●
● ●
●
●
●
● ●
● ●
●
● ● ●
●
● ●
●
● ●
● ●
● ●
● ●
●
●
● ●
●
●
● ● ● ●
● ●
● ●
●
●
●
●
●
● ● ● ● ●
● ●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ● ●
● ● ●
●
●
●
●
●
● ●
● ●
● ●
●
● ● ●
●
● ● ●
●
● ●
●
●
●
30 ●
● ●
●
● ●
60 ●
90 ● Atlantic Ocean
Extended Data Fig 3. bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
Chr 6 Chr 12 Chr 17 Chr 23 Vancouver Pacific Ocean Canada Autumn Canada Spring Atlantic Ocean Canada Spring Canada Autumn (NW) Canada Mixed Canada Autumn Norway Spring Atlantic Ocean Iceland Spring Greenland Summer (NE) Norway Spring Hudiksvall Spring Hästskär Spring Kalix Spring Karlskrona Spring Bornholm Autumn Fehmarn Autumn CentralBaltic Spring Gävle Autumn Baltic Sea Gävle Summer Gävle Spring Riga Autumn Riga Autumn Riga Spring Riga Spring Germany Spring Vaxholm Spring Gamleby Spring Kalmar Spring Rügen Spring Rügen Spring Schlei Spring Schlei Autumn Transition zone Träslövsläge Spring Ringköbing Spring Skagerrak Spring Kattegat Spring KattegatNorth Spring Hamburgsund Spring Gloppen Spring Lusterfjorden Spring Norwegian Lindås Spring Fjord Landvik Spring Landvik Spring Orkney Autumn CapeWrath Autumn Clyde Spring Ireland and TeelinBay Winter CapeWrath Spring Britain NorthSea Autumn Hebrides Mixed IsleOfMan Autumn CelticSea Winter Downs Winter
Extended Data Fig 4. NN SS bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
Chr 6 Chr 12 Chr 17 Chr 23 Pacific Ocean
Atlantic Ocean (NW)
Atlantic Ocean (NE)
Baltic Sea
Ireland and Britain
SS NN NS Extended Data Fig 5. bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint a (which was Westnot certified Atlantic by peer review) is the author/funder. All rights reserved.East No reuse Atlantic allowed without and permission.Baltic 65 70
60 65
55 Chr6 60
Latitude (°) 50 55 45
50 70 65 60 55 50 45 10 0 10 20 Longitude (°) b 65 70
60 65
55 60
Chr12 Latitude (°) 50 55 45
50 70 65 60 55 50 45 10 0 10 20 Longitude (°) c 65 70
60 65
Chr17 55 60
Latitude (°) 50 55
45 50 70 65 60 55 50 45 10 0 10 20 Longitude (°)
d 65 70
60 65
55 Chr23 60
Latitude (°) 50 55 45
50 70 65 60 55 50 45 10 0 10 20 Longitude (°)
Extended Data Fig 6. bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint a Chr(which 6 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 0.11%
0.13%
b 0.21% Chr 12
0.17%
c Chr 17 0.24%
0.31%
d Chr 23 0.16%
0.24%
Extended Data Fig 7. bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. a
b
Extended Data Fig 8. bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
North Atlantic Ocean
Baltic Sea
Transition zone
Norwegian Fjord
Ireland and Britain
Extended Data Fig 9 bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Riga Autumn Riga Autumn Riga Autumn Gävle Autumn BornholmBasin Autumn Schlei Autumn Fehmarn Kalix Spring Riga Spring Riga Spring Spring Vaxholm HastKar Spring Hudiksvall Spring Gävle Spring Gamleby Spring Kalmar Spring Karlskrona Spring Rügen Spring Rügen Spring CentralBaltic Spring Germany Spring Schlei Spring Gävle Summer Autumn IsleOfMan Autumn Orkney Autumn CapeWrath Autumn NorthSea Hebrides Mixed RingkobingFjord Spring Landvik Spring Landvik Spring Spring Traslovslage Kattegat Spring Hamburgsund Spring KattegatNorth Spring Skagerrak Spring Lindås Spring Lusterfjorden Spring Clyde Spring NSSH Spring Spring CapeWrath Gloppen Spring Norway Spring Iceland Spring Winter TeelinBay Downs Winter CelticSea Winter Autumn DalBoB Autumn DalGeB Autumn DalNsF Greenland Summer DalFB Mixed DalInB Spring DalNsS Spring DalInB Spring DalFB Mixed Greenland Summer DalNsF Autumn DalGeB Autumn DalBoB Autumn CelticSea Winter Downs Winter TeelinBay Winter Iceland Spring Norway Spring Gloppen Spring CapeWrath Spring NSSH Spring Clyde Spring Lusterfjorden Spring Lindås Spring Skagerrak Spring KattegatNorth Spring Hamburgsund Spring Kattegat Spring Traslovslage Spring Landvik Spring Landvik Spring RingkobingFjord Spring Hebrides Mixed NorthSea Autumn CapeWrath Autumn Orkney Autumn IsleOfMan Autumn Gävle Summer Schlei Spring Germany Spring CentralBaltic Spring Rügen Spring Rügen Spring Karlskrona Spring Kalmar Spring Gamleby Spring 0.06 Gävle Spring Hudiksvall Spring HastKar Spring 0.04 Vaxholm Spring Riga Spring Riga Spring 0.02 Kalix Spring Fehmarn Autumn Schlei Autumn 0.00 BornholmBasin Autumn Gävle Autumn Riga Autumn
Fig S1. bioRxiv preprint doi: https://doi.org/10.1101/2020.07.15.204214; this version posted July 15, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
Low differentiated Strongly differentiated Frequency 0e+00 4e+04 8e+04 0.0 0.02 0.08 0.1 0.2 0.3 0.4 Standard deviation of allele frequencies
Fig S2.