Deliverables
We examined the genetic structure among spring-spawning populations of Atlantic herring sampled at three locations in the Skagerrak and Norwegian spring-spawning herring sampled at two locations in the eastern North Sea using nine microsatellite DNA markers. The three Skagerrak locations were sampled in two consecutive years, 2002 and 2003, whereas Karmoy in the North Sea was sampled in 2002 and More further north off the norwegian coast, was sampled in 2003. Each sample consisted of ca 100 fish.
We found no evidence for genetic differentiation among Skagerrak samples, neither among locations nor between years. On the other hand was the More sample genetically differentiated from all Skagerrak samples and Karmoy from two of the three Skagerrak samples. These findings suggests a barrier to gene flow between Norwegian spring spawners and Skagerrak spring spawning stocks in spite of the known migration behaviour of Atlantic herring.
Development of embedded microsatellite marker in Atlantic herring major histocompatibility genes.
Joint scientific publication UWAG, UGOT and UHULL describing the development and population genetics of a polymorphic non-neutral microsatellite marker embedded in the MHC class II alpha gene of Atlantic herring. This marker adds to the low number of microsatellite markers available specifically for Atlantic herring to perform population genetics analyses.
Analysis of North Sea herring population structure using microsatellite molecular genetic markers.
In this study, we have examined genetic population structure across the major herring spawning aggregations in the North Sea and adjacent waters over two years, 2002 and 2003. We analysed 1660 spawning individuals across 9 microsatellite loci. Data were analysed using several approaches: taking into account the effect of location, year-class and sex, as well as pooling all individuals together, making no assumption as to the number of populations present in the dataset. Results suggest the presence of a genetically homogeneous unit off Northern Scotland, and a temporally stable pattern of isolation by distance determined predominantly by the divergence of the English Channel samples and, in 2003, by the Norwegian spring spawners.
Our data suggest that the current view of North Sea herring as a unit-stock might be adequate, but confirm the considerable degree of demographic independence of the herring populations in the English Channel. Despite major recent population collapses, genetic data indicated no evidence of bottlenecks affecting the genetic diversity of extant North Sea herring populations.
The work has focused on assessment of statistical power when using the statistical test procedures, sets of genetic markers, and sample sizes applied within the HERGEN project. To provide generality, however, we have also examined power for a wider set of tests, genetic markers, numbers of loci, sample sizes, and levels of genetic divergence.
Our analysis has followed three major lines, i.e. i) assessing the magnitude of actual power in some empirical and basic hypothetical settings, ii) comparing the efficiency of a few statistical approaches for testing for heterogeneity, and iii) evaluating in greater detail the phenomenon of a reduced power when combining multiple exact tests by means of Fisher's method, particularly in relation to chi-square.
The major observations may be summarized as follows. First, and as expected, power generally increases with the level of differentiation (FST), sample size, number of samples, and the number of loci and alleles, and uniform allele frequency distributions are better than skewed ones. Second, regardless of the statistical method employed, the power for detecting divergence may be substantial for frequently used sample sizes (say 50-100 individuals) and sets of genetic markers (say 5-20 allozyme or microsatellite loci) also at quite low values of FST (e.g. FST<0.01). For the herring microsatellites, for example, the probability of obtaining a significant result when sampling 50 fish from each of five populations approaches 100% for an FST of about 0.003, and the corresponding probability is around 50% for an FST as small as 0.001. Similarly, power may be high also for the less polymorphic allozymes when a reasonably large number of loci is scored (Figs. 1 and 2). Actually, in many studies using large numbers of highly polymorphic markers, the differences that are likely to be detected are so small that it may be questionable if they should be considered biologically meaningful for the issue at hand.
Third, we show that in some situations the choice of statistical method may be critical for detecting an existing genetic structure. Specifically, with a restricted number of alleles per locus summation of chi-square may outperform Fisher's exact and permutation tests, when the single locus P-values from these tests are combined by Fisher's method. This distinction between methods is most pronounced at skewed allele frequencies and few samples, and sometimes the already poor performance of the latter methods may decrease further with an increasing number of loci. Single or combined P-values from Fisher's exact test never display elevated a errors, though, and with large numbers of infrequent alleles the power of this approach may be higher than for chi-square.
Finally, an interesting observation refers to the good performance and the apparent robustness of the traditional chi-square test in situations when it is generally considered to be unreliable. On the other hand of the spectrum, the G-test (Sokal & Rohlf 1981) frequently tends to yield an unduly high proportion of false significances at loci segregating for low frequency alleles
On the basis of the present observations we recommend the following:
1. Assess power and α before launching a study. This can be done either through simulation, or more crudely through comparisons with estimates from similar investigations. For example, investigators should seriously consider whether the sample sizes and numbers of loci and alleles commonly used are really necessary for the particular question at hand.
2. When only considering a single locus (e.g. mtDNA) or contingency table, Fisher's exact test should be applied (rather than an approximation such as chi-square).
3. The G-test of Sokal and Rohlf (1981) should be avoided because of its tendency to produce excessive rates of false significances.
4. Be aware that combining exact P-values from multiple contingency tables (loci) by means of Fisher's method may under some circumstances result in a very low power. The risk for this phenomenon is most pronounced in pair-wise sample comparisons when dealing with di-allelic loci with skewed allele frequencies.
5. When exact tests combined with Fisher's method tend to yield an unduly low power, traditional chi-square may constitute a good alternative. Chi-square seems to be more robust than commonly appreciated. The risk for an inflated α error should still be considered, though, particularly in situations when comparing small samples (say n<20-30) with larger ones at loci with skewed allele frequencies.
6. In cases with several multi-allelic loci (e.g. 10 or more alleles), using Fisher's method to combine P-values obtained by Fisher's exact test seems to constitute the method of choice, because of a generally high power and an apparent tendency not to exceed the intended α level.
Genotyping of 11 polymorphic allozyme loci was performed using muscle, liver, and eye tissue samples of 2064 herring collected at 12 localities including the Baltic, Kattegat, Skagerrak, and North Seas. Sampling was repeated two consecutive years (2002 and 2003) for seven of these localities. Five of the 12 localities were included in a study of allozyme variability in herring that was carried out c. 20 years ago (Ryman et al. 1984, Heredity 53(3), 687-704), and the intention of including these localities in the present study was to create a possibility for long term temporal comparisons. Genotypic data from the Ryman et al. (1984) study is available and permit statistical comparisons between present day variability patterns and that observed two decades ago.
The sampling was not exclusively concentrated to spawning aggregations in the Ryman et al. study and only parts of the material from that study consisted of spawning individuals. In contrast, the sampling within HERGEN was concentrated to spawning aggregations. There were 1681 spawners out of the 2064 fish genotyped at allozyme loci. The observed levels of genetic variation, including expected heterozygosity and number of observed alleles in the present day material are comparable to that of Ryman et al. (1984). Similar to what was reported by Ryman et al. (1984), there are no indications of deviations from Hardy-Weinberg proportions in the present day material. Of a total of 407 individual tests for each sample and locus only two significant deviations were observed. No test remained significant when loci or populations were combined. Combining all samples into a single one resulted in genotype frequencies not significantly different from Hardy-Weinberg expectations.
Pairwise FST between sampling localities in 2002-2003 range from -0.003 to 0.003. Eight cases of significant differences between localities were observed, but none of these remain significant after Bonferroni correction. Overall spatial FST among present day localities is 0.0001 (P<0.001). No short-term genetic change could be observed when comparing samples collected in 2002 with those of 2003. The amount of spatial divergence appears to have remained consistent over the short and long term time periods examined; overall FST among the localities remain at c. 0.001 over the entire period. No long-term temporal differentiation can be observed at four of the five examined localities. At Fehmarn there is a statistically significant difference between the 1979 and 2002 collections, most likely reflecting sampling from different spawning stocks in 1979 vs. 2002.
The statistical power for detecting an FST of 0.005 is about 95% when using the present set of allozyme markers and sampling conditions (n = 100 and sampling from five populations). As we detect only limited occurrences of statistically significant heterogeneity it appears unlikely that the true degree of spatial or temporal divergence at the present allozyme markers is substantially larger than estimated in our study. Thus, the HERGEN sampling in 2002/03 has not changed the previous picture of very restricted spatial differentiation at allozyme loci in this species. The sampling of spawning aggregations has not affected the overall conclusion drawn twenty years ago.
Inter-calibration of microsatellite molecular genetic markers between partner institutes. Thirteen microsatellite loci were originally tested by the three partners involved in microsatellite DNA work (Partners 1a (DIFRES), 4 (UGOT), and 5 (UHULL)) on the same set of 40 standard individuals previously distributed among the 3 laboratories. Four of these were tetranucleotide loci developed for Atlantic herring by McPherson et al. (2001); eight were tetranucleotide loci developed for Pacific herring by Olsen et al. (2002), and one was a di-nucleotide locus isolated from Pacific Herring by O'Connell et al. (1998).
Different DNA extraction protocols (e.g., Phenol-Chloroform (Taggart et al. 1992), HotSHOT method (Truett et al. 2000), and Chelex resin (Walsh et al. 1991), as well as a broad array of different PCR amplification conditions were tested. DNA products were analysed using Pharmacia ALF-Express and MJ Research automated sequencing systems.
The following conclusions were reached:
- Chelex resin-based method was initially chosen as the extraction protocol, although this was later discarded in flavour of the HotSHOT method (Truett et al, 2000), since the former method yielded poorer quality of DNA. This allowed the PCR reaction to take place at more stringent conditions, so minimising the amplification of non-specific fragments.
- To ensure consistent genotyping and scoring across laboratories, single, unequivocal sizes were assigned to each allele in all loci.
- To ensure quality control of genotyping, two "standard heterozygote individuals" of known genotype were run alongside all samples, for each locus. The standards were chosen so as to cover a wide allelic range for each locus.
- To maximise consistency of genotyping, a procedure of "cross-check" scoring was adopted. The procedure involved the exchange between laboratories of 10 individuals to be scored blindly.
- Any "difficult-to-genotype" samples would not be scored in the first instance, but re-analysed and, if still "uncertain", ultimately recorded as missing values.
Use of these criteria permitted consistent scoring of alleles using 3 types of automatic sequencers situated in the 3 different laboratories, and would provide a valuable first step for other projects where microsatellite data from multiple sources is to be reliably combined.
McPherson et al. 2001. Molec. Ecol. Notes, 1: 31-32; O'Connell et al. 1998. Molec. Ecol. 7: 357-363; Olsen et al. 2002. Molec. Ecol. Notes, 2: 101-103; Taggart et al. 1992. J. Fish Biology, 40: 963-965; Truett et al, 2000 BioTechniques, 29: 52-54; Walsh et al. 1991. Biotechniques, 10: 506-513.
For stock assessment purposes different metrics have been applied for separating different stock components in mixed samples. Earlier population based methods using counts of keeled scales and vertebral series have been replaced by individual differentiation using otolith microstructure formed during the larval phase.
The work constituting the present HERGEN result "Phenotypic and Otolith Characterization - 29283" focused on collation and analysis of a number of phenotypic characters and additional development of new otolith based methods to identify spawning populations in mixed samples. A number of growth and maturation parameters differed between spawning components and e.g. mean vertebral counts were found to be different between some populations with the same spawning period.
The performance of the otolith microstructure method was tested and further developed within the project. High reproducibility of the visual inspection method was demonstrated in blind readings of spring and autumn/winter spawning populations. Experienced readers had average misclassification rates of about 2% and between reader differences were about 1%.
An objective method of daily increment width analysis was developed aiming at identification of strayers not spawning in the same season that they were hatched. Otolith silhouette shape was analysed by Elliptic Fourier Transformation (EFT) for a number of spawning populations from all seasons and major sea areas. Cross-validated discriminant analysis of selected EFT components showed classification success ranging from 64-84%. Similarity by Euclidean distance between average population shapes indicated an environmental impact of overlap in summer feeding area.
In general the otolith microstructure methods come out with very precise estimates of hatching period at the individual level, whereas they have limited capability of identifying different populations with the same spawning time. Shape analysis performs better at the population level but is less precise at the individual level. We may therefore conclude that otolith shape in combination with larval microstructure provides an effective additional tool for identification of herring population structure and that as a base line is readily available for splitting mixed stock catches of mature herring into management units. As such they are excellent supplements to microsatellite analysis of mixed stocks.
The prevailing hypothesis underpinning management decisions for herring in the North Sea, Skagerrak, Kattegat and Western Baltic is that spawning time characteristics are sufficient for separating the two major components, North Sea autumn spawners and Western Baltic spring spawners. The combined results on genetics and phenotype in the present project demonstrate strong population structuring following the salinity clines from the North Sea towards the Western Baltic. This structuring indicates important genetic diversity per se but may also help understanding the complex variation in spatial and temporal abundance created by differential migration pattern of the different population subunits.
It is predicted that the combination of the entire range of phenotypic and genetic methods may provide detailed information of population composition in mixed stocks even outside the range of populations present in the base line. Even when retaining the present management units a finer scale resolution of population structure will improve the stock assessment procedure for the area.
Genetic structure among herring populations from the Skagerrak to the western Baltic (Rugen) was examined using analysis of nine microsatellite markers. From each of ten spawning locations sampled across two consecutive years we genotyped approximately 200 fish, totalling genetic analysis of 1951 fish. Statistical analyses of allele frequency differences indicated significant population structure among most of the pair-wise samples, and that the differences were stable across the two years. Levels of genetic differentiation followed an isolation-by-distance model, signifying that herring originating from more distant locations also exhibited larger genetic differences.
The examined area spans the transition zone between the deep and highly saline North Sea and the more shallow and brackish Baltic sea, and partial regression analysis incorporating information for genetic, geographic and environmental differences among spawning population indicated that difference in salinity levels on spawning sites was highly correlated with genetic differences, irrespective of geographic distances among populations. This can be interpreted as an indication that genetic differences among herring populations are maintained by salinity-correlated selective differences on spawning sites. Although salinity is known to be a strong selective factor in other organisms, direct evidence for salinity acting as a selective factor shaping population structure in herring is still needed. The implication of the study is that herring of different geographical origin, in spite of their extensive migratory behaviour, show reproductive isolation and that recruitment to a large extent is locally determined.
Herring fisheries commonly exploit aggregations composed of individuals of mixed population origin. We used microsatellite DNA information obtained both in samples collected on spawning sites and in samples of mixed stocks to estimate relative population compositions in mixed aggregations in the northern North Sea and the Skagerrak. We developed a tool to determine mixed stock compositions by applying a bayesian mixed-stock analysis algorithm.
Our analysis method identified hitherto unknown spatially and temporally explicit migratory patterns of mixed aggregations of herring in the Skagerrak. Aggregations sampled in the northern North Sea showed a strong signal of major contributions from populations originating from the western North Sea and English Channel. Aggregations in the Skagerrak exhibited substantial mixing of fish originating in the North Sea, Skagerrak, Inner Danish waters and the western Baltic. We also showed that the statistical power associated with our analysis framework was high and thus that the newly developed method presents a powerful tool in fisheries assessment applications.
Atlantic herring (Clupea harengus) has traditionally been divided into different stocks based on spawning behaviour and meristic or morphological characters. It is unclear, however, if these stocks represent reproductively distinct populations. Recently, the application of new molecular methods has proven useful to distinguish between populations of marine fish. Here, we compare the ability of four different genetic markers, allozymes, mitochondrial DNA, nuclear microsatellite DNA and immune defence genes (MHC) to discriminate between stocks of herring. MtDNA and microsatellites are supposedly selectively neutral, whereas allozymes and especially MHC may be under natural selection. Using the same individual fish for all genetic and morphological analyses, we compare samples (ca75 fish) collected at five different sites in two subsequent years in the Baltic, Skagerrak and the North Sea.
Microsatellites and MtDNA were the most powerful markers to detect differentiation among samples. For allozymes only one single sample was distinct from the rest, Tjome 2002, whereas MHC analysis indicated that the Rugen samples were differentiated from the rest. Microsatellites and to some extent MHC showed higher differentiation among localities than between temporally spaced samples indicating temporal stability of the population structure. Concordance among marker types appeared to be fairly low. This observation could be due to several factors that differ between marker types, such as levels of heterozygosity, allellic richness, mutation rates and susceptability to selective forces. The two marker types that potentially could be under selection, MHC and allozyme loci, showed distinct differentiation for certain samples. However, if this pattern could be linked to specific environmental conditions is presently unknown.
The population structure suggested by microsatellite data coincides with the geographical separation of Baltic, Skagerrak and North Sea locations. This finding together with the fact that microsatellite population structure was highly significant statistically, suggests that microsatellites are the most powerful genetic marker, among the ones investigated here, to distinguish between different stocks of herring.
The result 29273 is a series of recommendations that can be taken forward to monitor levels of biodiversity of different components of herring in the fisheries in the North Sea and into ICES Divisions IIIa and IIIb and the western Baltic. Other HERGEN results informed the decisions taken here and led to the conclusion that two different approaches were required to monitor biodiversity in the herring fisheries in the area of study.
Genetics showed little apparent structuring of the North Sea fished population but otolith microstructure allowed the Downs winter spawning component to be separated off and therefore monitored. It has been recommended that 100 samples (2500 fish) should be taken throughout the fishery (as well as through surveys) annually to determine proportions of Downs herring within the main body of the North Sea fishery and therefore inform the assessment and management process. In IIIa, b, and western Baltic a combination of molecular genetic techniques and otolith microstructure gave the most cost effective method for determining mixtures of components within the fishery and therefore for monitoring biodiversity.
Until now, spring spawning components in the fishery had been unable to be determined beyond "western Baltic spring spawners" when it was thought highly likely that there were other spring spawning components within the area. It has been recommended that 30 samples (3000 fish) should be taken throughout the fishery annually to determine relative proportions. This project therefore provides methodology enabling their discrimination and provides the recommendation for the levels of sampling required for monitoring of the proportions of the various components.