Limits...
Ecological niche models and coalescent analysis of gene flow support recent allopatric isolation of parasitoid wasp populations in the Mediterranean.

Lozier JD, Mills NJ - PLoS ONE (2009)

Bottom Line: The coalescent analysis revealed low levels of migration among western and eastern Mediterranean populations (Nm<1) that were not statistically distinguishable from zero.Overall, environmental characteristics were similar among the population clusters, though significant differences did emerge.These results support contemporary allopatric isolation of Mediterranean populations of A. transcaspicus, which together with previous analyses indicating partial behaviorally mediated reproductive isolation, suggest that the early stages of cryptic speciation may be in progress.

View Article: PubMed Central - PubMed

Affiliation: Department of Entomology, University of Illinois, Urbana, Illinois, USA. jdlozier@life.uiuc.edu

ABSTRACT

Background: The integration of multiple complementary approaches is a powerful way to understand the processes of diversification and speciation. The parasitoid wasp Aphidius transcaspicus Telenga (Hymenoptera: Braconidae) is a parasitoid of Hyalopterus aphids across a wide geographic range. This species shows a remarkable degree of genetic structure among western, central, and eastern Mediterranean population clusters. In this paper we attempt to better characterize this genetic structure.

Methodology/principal findings: We use a Bayesian coalescent analysis of gene flow under the Isolation with Migration model using mitochondrial and microsatellite markers together with climate-based ecological niche models to better understand the genetic structure of A. transcaspicus in the Mediterranean. The coalescent analysis revealed low levels of migration among western and eastern Mediterranean populations (Nm<1) that were not statistically distinguishable from zero. Niche models showed that localities within population clusters each occupy areas of continuously high environmental suitability, but are separated from each other by large regions of completely unsuitable habitat that could limit dispersal. Overall, environmental characteristics were similar among the population clusters, though significant differences did emerge.

Conclusions/significance: These results support contemporary allopatric isolation of Mediterranean populations of A. transcaspicus, which together with previous analyses indicating partial behaviorally mediated reproductive isolation, suggest that the early stages of cryptic speciation may be in progress.

Show MeSH
Average logistic probability of presence (±1 SD) for localities in K1 (n = 30), K2 (n = 5), and K3 (n = 28) population clusters in each of the examined ENMs.
© Copyright Policy
Related In: Results  -  Collection


getmorefigures.php?uid=PMC2691581&req=5

pone-0005901-g003: Average logistic probability of presence (±1 SD) for localities in K1 (n = 30), K2 (n = 5), and K3 (n = 28) population clusters in each of the examined ENMs.

Mentions: The reduced ENMs developed using localities from either K1 (average training and test AUC values of 0.994±0.001 SD and 0.992±0.002 SD, respectively) or K3 (average training and test AUC values of 0.981±0.004 SD and 0.982±0.018 SD, respectively) alone also performed well in predicting the range-wide distribution of A. transcaspicus (Figure 2). Each of these predictions is broadly similar in scope to the full ENM, particularly the K1-only ENM. As in the full model, both reduced models predicted a coastal restriction for A. transcaspicus and also identified breaks in bioclimatic suitability found along the central Libyan coast and along the northern coast of the Adriatic Sea. Overall this suggests that there is some level of niche similarity between the genetic clusters, though there are several important differences. For the K1-only ENM none of the localities in K1 or K2 were omitted completely from the predicted distribution. Most of the K3 localities were also predicted successfully, with the exception of 5 localities in Israel, the Turkey locality, and a single locality from Pilion-Magnisia, Greece that was georeferenced from [30]. The omissions from the K3-only model were more pronounced, with regions containing the Turkey and Pilion-Magnisia [30] localities (K3), the three Italian localities (K2), and the southernmost Tunisian and all eastern Spanish localities (K1) classified as unsuitable. Comparison of logistic probabilities of occurrence among clusters and models shows clear declines for localities not included in reduced models relative to the full model (Figure 3). While, on average all sampled localities were predicted with high probability in the full ENM, in the K1-only ENM, the K2 and K3 localities were dramatically reduced, and likewise, for the K3-only ENM the K1 and K2 localities had a low probability of occurrence (Figure 3). This suggests that localities in each cluster contain unique bioclimatic niche information and thus a model generated using data from only a single cluster does a poor job of predicting the detailed geographic distributions of parasitoids in the other clusters, even though, at least qualitatively, the K1-only ENM does a fairly good job of reproducing the distribution map from the full ENM.


Ecological niche models and coalescent analysis of gene flow support recent allopatric isolation of parasitoid wasp populations in the Mediterranean.

Lozier JD, Mills NJ - PLoS ONE (2009)

Average logistic probability of presence (±1 SD) for localities in K1 (n = 30), K2 (n = 5), and K3 (n = 28) population clusters in each of the examined ENMs.
© Copyright Policy
Related In: Results  -  Collection

Show All Figures
getmorefigures.php?uid=PMC2691581&req=5

pone-0005901-g003: Average logistic probability of presence (±1 SD) for localities in K1 (n = 30), K2 (n = 5), and K3 (n = 28) population clusters in each of the examined ENMs.
Mentions: The reduced ENMs developed using localities from either K1 (average training and test AUC values of 0.994±0.001 SD and 0.992±0.002 SD, respectively) or K3 (average training and test AUC values of 0.981±0.004 SD and 0.982±0.018 SD, respectively) alone also performed well in predicting the range-wide distribution of A. transcaspicus (Figure 2). Each of these predictions is broadly similar in scope to the full ENM, particularly the K1-only ENM. As in the full model, both reduced models predicted a coastal restriction for A. transcaspicus and also identified breaks in bioclimatic suitability found along the central Libyan coast and along the northern coast of the Adriatic Sea. Overall this suggests that there is some level of niche similarity between the genetic clusters, though there are several important differences. For the K1-only ENM none of the localities in K1 or K2 were omitted completely from the predicted distribution. Most of the K3 localities were also predicted successfully, with the exception of 5 localities in Israel, the Turkey locality, and a single locality from Pilion-Magnisia, Greece that was georeferenced from [30]. The omissions from the K3-only model were more pronounced, with regions containing the Turkey and Pilion-Magnisia [30] localities (K3), the three Italian localities (K2), and the southernmost Tunisian and all eastern Spanish localities (K1) classified as unsuitable. Comparison of logistic probabilities of occurrence among clusters and models shows clear declines for localities not included in reduced models relative to the full model (Figure 3). While, on average all sampled localities were predicted with high probability in the full ENM, in the K1-only ENM, the K2 and K3 localities were dramatically reduced, and likewise, for the K3-only ENM the K1 and K2 localities had a low probability of occurrence (Figure 3). This suggests that localities in each cluster contain unique bioclimatic niche information and thus a model generated using data from only a single cluster does a poor job of predicting the detailed geographic distributions of parasitoids in the other clusters, even though, at least qualitatively, the K1-only ENM does a fairly good job of reproducing the distribution map from the full ENM.

Bottom Line: The coalescent analysis revealed low levels of migration among western and eastern Mediterranean populations (Nm<1) that were not statistically distinguishable from zero.Overall, environmental characteristics were similar among the population clusters, though significant differences did emerge.These results support contemporary allopatric isolation of Mediterranean populations of A. transcaspicus, which together with previous analyses indicating partial behaviorally mediated reproductive isolation, suggest that the early stages of cryptic speciation may be in progress.

View Article: PubMed Central - PubMed

Affiliation: Department of Entomology, University of Illinois, Urbana, Illinois, USA. jdlozier@life.uiuc.edu

ABSTRACT

Background: The integration of multiple complementary approaches is a powerful way to understand the processes of diversification and speciation. The parasitoid wasp Aphidius transcaspicus Telenga (Hymenoptera: Braconidae) is a parasitoid of Hyalopterus aphids across a wide geographic range. This species shows a remarkable degree of genetic structure among western, central, and eastern Mediterranean population clusters. In this paper we attempt to better characterize this genetic structure.

Methodology/principal findings: We use a Bayesian coalescent analysis of gene flow under the Isolation with Migration model using mitochondrial and microsatellite markers together with climate-based ecological niche models to better understand the genetic structure of A. transcaspicus in the Mediterranean. The coalescent analysis revealed low levels of migration among western and eastern Mediterranean populations (Nm<1) that were not statistically distinguishable from zero. Niche models showed that localities within population clusters each occupy areas of continuously high environmental suitability, but are separated from each other by large regions of completely unsuitable habitat that could limit dispersal. Overall, environmental characteristics were similar among the population clusters, though significant differences did emerge.

Conclusions/significance: These results support contemporary allopatric isolation of Mediterranean populations of A. transcaspicus, which together with previous analyses indicating partial behaviorally mediated reproductive isolation, suggest that the early stages of cryptic speciation may be in progress.

Show MeSH