Genome sequencing reveals hidden exceptions in toothfish connectivity

Whole-genome resequencing and reduced-representation sequencing of Antarctic toothfish, Dissostichus mawsoni, found no clear circumpolar population structure, but uncovered exceptions among older and younger cohorts: genetic diversity declined and admixture increased in the Ross Sea between fish born before 2000 and those born afterward. The study sampled 108 fish from the Weddell Sea with 3RADseq and 41 fish from three Southern Ocean management areas with whole-genome resequencing, revealing a more complex history than the prevailing view that the species forms one panmictic metapopulation. The results do not establish a circumpolar decline, because only one pre-2000 fish was available from the Weddell Sea and one from East Antarctica, but they identify temporal variation and elevated inbreeding in the Ross Sea as important features that conventional geographic connectivity assessments may miss.

Antarctic toothfish is a long-lived, slow-developing predator that can reach about 2 meters and does not generally mature until 12–15 years of age. Its life cycle links offshore spawning, sea-ice development, transport through the Southern Ocean, and feeding on the continental shelf. This dependence on currents can maintain gene flow among distant populations, while the species’ reliance on sea ice and its exposure to an intensive, geographically uneven fishery introduce other forces that may alter that connectivity. The fishery expanded from the Ross Sea to East Antarctica and the Weddell Sea, with particularly high legal catch in the Ross Sea after the late 1990s and substantial illegal, unreported, and unregulated fishing during earlier decades. These features make toothfish an unusual test of whether widespread oceanic dispersal is sufficient to keep an exploited population genetically uniform.

Previous genetic studies reached conflicting conclusions. Some reported differentiation among the Weddell Sea, Ross Sea, and East Antarctic populations, whereas others found little or no restriction to gene flow. A central limitation was that most studies compared spatial samples without explicitly separating cohorts. Cohort separation matters because broadcast-spawning populations can display “chaotic genetic patchiness”: differences in recruitment and the proportions of successful breeders can change the genetic composition of a population from one cohort to another. A population that appears panmictic in one sampling period may therefore be partly isolated in another. To address this possibility, the researchers combined 3RADseq data from 108 Weddell Sea fish with whole-genome resequencing data from 41 toothfish collected across Commission for the Conservation of Antarctic Marine Living Resources management areas 48, 58, and 88.

The two genetic approaches differed substantially in genomic representation. The 3RADseq method generated 21,002 final SNPs from about 1% of the genome, whereas the whole-genome data produced roughly 1.9–5.4 million filtered SNPs, depending on sequencing coverage and analysis, with millions remaining after linkage-disequilibrium pruning. Because the full-genome libraries had uneven coverage of 1.13-fold to 27.09-fold, investigators created 2-fold and 5-fold downsampled datasets and also retained an undownsampled set. They used genotype likelihoods for much of the population analysis, allowing uncertainty from low coverage to be incorporated, and called genotypes only in the 5-fold dataset. Birth cohorts were estimated from total length, but only to within five years, so the study grouped fish broadly as born before or after 2000, a boundary chosen because Commission catch documentation, surveillance, and management reforms sharply reduced illegal fishing around that time.

Neither principal component analysis nor admixture analysis detected geographic population structure among Weddell Sea subareas. Whole-genome results likewise failed to align genetic clusters with the three management areas, supporting the broad panmixia hypothesis. However, grouping samples by cohort exposed a different pattern. Most pre-2000 fish showed little admixture, whereas many post-2000 fish were strongly admixed, apart from five Ross Sea individuals and a single pre-2000 exception. Weighted F_ST values across geographic, cohort, coverage, and sex groupings ranged from 0.0218 to 0.0487, but did not isolate a clear signal attributable to management area or cohort. The investigators cautioned that technical effects, including sequencing depth, reference-mapping bias, batch composition, and low-coverage uncertainty, could contribute to cohort patterns in a dataset containing fewer than 50 whole-genome libraries.

Genetic-diversity estimates showed a marked regional contrast. In the Weddell Sea, expected heterozygosity was higher in management subarea 48.4 than in subareas 48.1 and 48.2. Across the three circumpolar management areas, overall differences in heterozygosity were not statistically significant, although the Ross Sea showed a visible trend toward lower diversity and lower Watterson’s theta. The clearest temporal comparison was possible only in the Ross Sea: heterozygosity fell significantly between the pre-2000 and post-2000 cohorts, a result reproduced in both the 2-fold and undownsampled datasets. Watterson’s theta showed the same decline. By contrast, cohort comparisons in the Weddell Sea and East Antarctica could not be calculated because each contained only one pre-2000 individual. The evidence therefore supports reduced diversity over time within the Ross Sea sample, not a demonstrated decline across all toothfish populations.

Inbreeding measures showed the Ross Sea as the region of greatest concern. Individual inbreeding coefficients were significantly higher there than in the Weddell Sea, and runs of homozygosity produced the same regional pattern in the 5-fold genotype dataset. Neither measure, however, differed significantly between the Ross Sea cohorts, so the analyses do not demonstrate that inbreeding itself increased after 2000. Tajima’s D was negative in all three management areas and tended to be lower in the Ross Sea, while post-2000 Ross Sea fish showed a tendency toward higher values than their pre-2000 counterparts. Effective population size, inferred with SMC++, varied little over recent time, with the Ross Sea retaining the highest estimate among the areas and the Weddell Sea showing a recent decline. The authors stressed that effective population size is difficult to infer in large marine populations, that the model tends to smooth abrupt changes, and that four fish per area constrained the demographic analysis.

Several mechanisms could explain the combination of stronger admixture and lower diversity in younger Ross Sea fish. One possibility is that incoming source populations carried less genetic variation than the resident population, so gene flow could increase mixture while reducing overall heterozygosity. Recruitment variability is another candidate: toothfish recruitment is linked to sea-ice drift and climate variability, including the El Niño–Southern Oscillation, and anomalous recruitment years could alter the diversity of successive cohorts. Fishing may also have contributed. Historical illegal fishing cannot fully explain the particularly strong Ross Sea pattern because the greatest illegal catch occurred elsewhere, but sustained legal exploitation in the Ross Sea may have affected the genetic composition of survivors and subsequent recruits. The study does not distinguish conclusively among these explanations, and its observations are not evidence that fishing caused the reported genomic changes on their own.

For fisheries management, the main implication is that spatial surveys of connectivity may be insufficient. Toothfish otoliths are routinely collected to determine age, and the study demonstrated that DNA remaining on their surfaces can support whole-genome library preparation. This creates the possibility of repeatedly monitoring genetic diversity using samples already obtained by the fishery. The researchers proposed standardized annual monitoring and larger, more balanced samples from every region and cohort, especially because apparent stability across a broad Southern Ocean distribution could conceal cohort-specific contractions, inbreeding, or altered gene flow. The study also identified more than 121,000 putatively adaptive SNPs with one outlier-detection method, although other methods did not identify outliers and the adaptive status of these loci remains to be tested. For toothfish, whose early development depends on sea ice that has undergone pronounced recent changes, resolving how climate, recruitment, fishing, and genome-wide diversity interact will be necessary to distinguish a genuinely panmictic population from temporary and regionally specific departures from that condition.

Subject of Research: Genetic connectivity, diversity, and historical change in exploited Antarctic toothfish

Article Title: Whole‐Genome Resequencing Reveals Exceptions to Panmixia in the Exploited Top Fish Predator Antarctic Toothfish (Dissostichus mawsoni)

Article References: Caccavo, J. A., Arantes, L. S., Celemín, E., Mbedi, S., Sparmann, S., & Mazzoni, C. J. (2026). Whole‐Genome Resequencing Reveals Exceptions to Panmixia in the Exploited Top Fish Predator Antarctic Toothfish ( Dissostichus mawsoni ). Ecology and Evolution, 16(10), Article e74262. https://doi.org/10.1002/ece3.74262

Image Credits: AI Generated

DOI: 10.1002/ece3.74262

Keywords: Antarctic toothfish, whole-genome resequencing, population genetics, panmixia, genetic diversity, Ross Sea, fisheries management, Southern Ocean, chaotic genetic patchiness, inbreeding, Whole, Genome