The Atlantic bluefin tuna is a highly migratory species emblematic of the challenges associated with shared fisheries management. In an effort to resolve the species’ stock dynamics, a genome-wide search for spatially informative single nucleotide polymorphisms (SNPs) was undertaken, by way of sequencing reduced representation libraries. An allele frequency approach to SNP discovery was used, combining the data of 555 larvae and young-of-the-year (LYOY) into pools representing major geographical areas and mapping against a newly assembled genomic reference. From a set of 184,895 candidate loci, 384 were selected for validation using 167 LYOY. A highly discriminatory genotyping panel of 95 SNPs was ultimately developed by selecting loci with the most pronounced differences between western Atlantic and Mediterranean Sea LYOY. The panel was evaluated by genotyping a different set of LYOY (n= 326) and from these 77.8% and 82.1% were correctly assigned to western Atlantic and Mediterranean Sea origins, respectively. The panel revealed temporally persistent differentiation among LYOY from the western Atlantic and Mediterranean Sea (FST = 0.008, p=0.034). The composition of six mixed feeding aggregations in the Atlantic Ocean and Mediterranean Sea was characterized using genotypes from medium (n = 184) and large (n = 48) adults, applying population assignment and mixture analyses. The results provide evidence of persistent population structuring across broad geographic areas and extensive mixing in the Atlantic Ocean, particularly in the mid-Atlantic Bight and Gulf of St. Lawrence. The genomic reference and genotyping tools presented here constitute novel resources useful for future research and conservation efforts.

[Usage Notes] Puncher et al_Thunnus thynnus_95loci_723individuals_27samples Puncher et al_Thunnus thynnus_221loci_167individuals_9samples Puncher et al_Thunnus thynnus_384 SNPs with flanking sequences "Puncher et al_Thunnus thynnus_384 SNPs with flanking sequences.txt" contains the flanking sequences of 384 SNPs discovered using reduced representation sequence libraries, 221 of which passed quality criteria and 95 that were used for population structure analysis.

Peer reviewed