Meta-analysis of factors influencing population differentiation in yellowfin tuna (Thunnus albacares)

Decades of study have attempted to describe the population genetic structure of tuna species, including a significant number of studies focused on yellowfin (Thunnus albacares). Very often, analyses do not agree on how many stocks exist per species, and where their boundaries lie. This is possibly because studies vary so much across numerous variables, such as the geographic range covered in a study, the number and type of genetic markers used, and the number of sites sampled. This meta-analysis of 22 yellowfin studies attempts to standardize and isolate the key variables to assessing the strength of correlation with the resulting number of populations a study observes. Overall trends across the studies suggest that genetic markers with a high probability of being under selective constraints, or are located in coding regions, are more or less likely to sense population structure depending on geographic range coverage. Alternatively, when assessing neutral or non-coding genomic regions, studies also benefit from (i) more polymorphic and numerous loci, and (ii) increasing the number of sampling locations analyzed, both of which increase the statistical power of an analysis. Finally, trends were clearest when groups of accurately identified coding and non-coding, or neutral and non-neutral, studies were further subdivided by whether they used mitochondrial or nuclear DNA, confirming that analyses of the two genomes should not be compared directly. Our meta-analysis provides concrete support for long held assumptions about the relationships between population genetics study parameters and outcomes, and provides guidance for future studies on how to maximize the likelihood of identifying population structure where it exists.