Assessing population structure in the face of isolation by distance: Are we neglecting the problem?

Abstract

AimTo investigate how the potential bias from isolation by distance (IBD) in inferences of population structure has been addressed, in studies aiming biodiversity conservation.LocationGlobal.MethodsWe reviewed the literature on the impact of IBD on the performance of the widely used software structure. We also performed a literature survey in the Web of Knowledge to assess how data have been analysed in biodiversity conservation studies when IBD is detected.ResultsBy reviewing the literature on structure performance in IBD data sets, we found a high number of studies showing that structure outputs are extremely affected by IBD. This misleading inference results principally in the detection of artificial genetic clusters. The literature survey showed that IBD was present in most data sets (60.56%) and that a substantial number of the articles only tested IBD by Mantel tests (82.21%). The most concerning result is that several articles have been using structure even after detecting IBD (57.99%), and a substantial number of them are drawing formal conservation strategies, notwithstanding the potentially biased results (51.49%).Main conclusionsOur results are of great concern, as conservation strategies may be distinct under different population structure, and its success can be affected by the incorrect identification of populations. We recommend possible stages to be considered in the presence of IBD, which can help conservation investigators before the proposition of explicit conservation strategies.

Conserving on the edge: genetic variation and structure in northern populations of the endangered plant Dracocephalum ruyschiana L. (Lamiaceae)

“…This is comparable to what was found in the close relative D. austriacum in the Czech and Slovak Republics; high genetic diversity within populations (80%) and relatively low differentiation among populations both within and between regions (Dostálek et al 2010). In D. ruyschiana, isolation by distance was observed which may bias the Structure results (see Perez et al 2018 and references therein). However, the spatial multivariate analysis (MULTISPATI-PCA) takes spatial data into account and the result obtained is very similar to the Structure result.…”

Ancestral and neo-sex chromosomes contribute to population divergence in a dioecious plant

“…Clustering of populations in such analyses as those performed above can be caused by geographic discontinuity in sampling (Balkenhol et al 2015; Bradburd et al 2016; Perez et al 2018). In such cases, clustering may be a product of the interaction between incomplete sampling and isolation-by-distance rather than discrete barriers to gene flow.…”

Ancestral and neo‐sex chromosomes contribute to population divergence in a dioecious plant

“…As our sampling design did not include the purported cytotype contact zone (shaded area in Fig. 1C; Smith 1969), and thus created a discontinuity in sampling, clustering may be a product of the interaction between incomplete sampling and isolation-bydistance rather than representing a discrete barrier to gene flow (see Balkenhol et al 2015;Bradburd et al 2016;Perez et al 2018). To evaluate this possibility, we examined the estimated effective migration surface (EEMS; Petkova et al 2016), which is a Bayesian estimate of effective migration rates based on SNP frequencies and a prior hypothesis of isolation-by-distance.…”