The K = 2 conundrum

Abstract

Assessments of population genetic structure have become an increasing focus as they can provide valuable insight into patterns of migration and gene flow. STRUC-TURE, the most highly cited of several clustering-based methods, was developed to provide robust estimates without the need for populations to be determined a priori. STRUCTURE introduces the problem of selecting the optimal number of clusters, and as a result, the DK method was proposed to assist in the identification of the "true" number of clusters. In our review of 1,264 studies using STRUCTURE to explore population subdivision, studies that used DK were more likely to identify K = 2 (54%, 443/822) than studies that did not use DK (21%, 82/386). A troubling finding was that very few studies performed the hierarchical analysis recommended by the authors of both DK and STRUCTURE to fully explore population subdivision. Furthermore, extensions of earlier simulations indicate that, with a representative number of markers, DK frequently identifies K = 2 as the top level of hierarchical structure, even when more subpopulations are present. This review suggests that many studies may have been over-or underestimating population genetic structure; both scenarios have serious consequences, particularly with respect to conservation and management. We recommend publication standards for population structure results so that readers can assess the implications of the results given their own understanding of the species biology.

Comparative phylogeography of trans‐Andean freshwater fishes based on genome‐wide nuclear and mitochondrial markers

“…For all species, we found the optimal K to be 2 based on the Δ K statistic analysed with the unlinked nDNA data set (Supporting Information Figure , Evanno et al, ). Although there have been recent concerns that structure analyses tend to be biased in favour of K = 2 (Janes et al, ), we note that for most species our K schemes appear to be consistent with the results inferred using PCoA (Supporting Information Figure ). These results largely match the major clades obtained with the phylogeographic inference and concordance factors, exhibiting geographic boundaries of inferred genetic clusters that are largely concordant among species in three major geographic units: (a) CAU, MAG, and SJ (Magdalena system) for M. muyscorum, E. humboldtii, S. aequilabiatus and A. pardalis (Figure b–e); (b) ATR + TUI for E. humboldtii (Figure d); and (c) ATR + SIN for M. muyscorum (Figure b).…”

Range limits in sympatric cryptic species: a case study in Tetramorium pavement ants (Hymenoptera: Formicidae) across a biogeographical boundary

“…Therefore, Meirmans (2015) suggested discussing all clustering results that warrant a biological interpretation because clustering analysis is an exploratory analysis with interpretations at multiple levels. Finally, a problem commonly raised in the literature is the overrepresentation of the K = 2 situation when using DK approaches (Gilbert, 2016;Janes et al, 2017). This is all the more appropriate here as the clusters are of unequal sizes, a situation where it is more difficult to accurately estimate K across an existing population structure (Puechmaille, 2016).…”

Demographic modelling reveals a history of divergence with gene flow for a glacially tied stonefly in a changing post‐Pleistocene landscape

“…Benestan et al., ), but it's still an understudied finding in a field dominated by clear expectations and results (but see Janes et al. discussion of studies underestimating K after an initial and clear K = 2 result). Specifically, our aim was to better understand how the full 6,819 SNP data set could give such opposing results when analysed via DAPC ( K = 1) or Admixture ( K = 12).…”