The most widespread genetic clusters were the Red and Pink clusters found within thirteen and eight populations along a distance of 130 km from Kirchentellinsfurt to Augsburg and from Tübingen to Großaitingen, respectively.
Population structure was detected with the software STRUCTURE 2.1 [ 81] (without prior population information), which uses a Markov chain Monte Carlo (MCMC) algorithm to define the most likely genetic clusters on the basis of multilocus genotype data.
In this analysis, populations were reduced to individuals from the single, most dominant genetic cluster in that habitat (as indicated by the STRUCTURE results).
Additionally, if k is allowed to vary, it is apparent that fish from Lake Apoyo form the most distinct genetic cluster, separate from fish in the other lakes at k = 2. Following this, at k = 3 the fish from Lake Xiloa are found to be distinct, whereas fish from Lake Masaya tend to group strongly with those from Lake Nicaragua.
Accordingly with the high genetic differentiation and the low proportion of shared diversity among unstocked lakes, most of the genetic clusters delineated by structure corresponded to a unique lake in the case of unstocked populations.
The fully Bayesian model-based clustering method implemented in STRUCTURE v2.2 [21] was used to infer the most likely number of genetic clusters (K) in the data set.
The most likely number of genetic clusters (K) was inferred using the method of Evanno et al. [87], where both the log probability and the rate of change in log probability are considered.
As this coincides with one of the optima in ΔK (Fig. 2B), we have interpreted these results as 20 being the most likely number of genetic clusters in the data, although K = 2 and K = 7, the other two optima in ΔK, were also explored (discussed below but data not shown).
In 2005, the most likely number of genetic clusters was 12 (including the LP cluster).
We concluded that K = 2 is the most likely number of genetic clusters.
