This action funds an NSF Postdoctoral Research Fellowship for FY 2010. The fellowship supports a research and training plan entitled "Landscape genetic modeling of environmentally associated morphological and genetic variation" for Ian Wang. The host institution for this research is Harvard University, and the sponsoring scientist is Jonathan Losos.
To date, the vast majority of landscape genetics research has focused on the relationship between population genetic structure and geographic distance, without considering the potential effects of environmental variation. Structural equation modeling (SEM), which utilizes a series of regression and model-fitting analyses to test hypotheses about the relationships between multiple variables, provides a logical framework for examining associations between geographic distance, ecological differences, and genetic structure. This research utilizes a computationally intensive SEM analysis and data acquired from five major DNA, GIS, and museum record databases to test whether patterns of genetic structure in 12 species of Anolis lizards are associated with between-population geographic distance or ecological differences. The implementation of this method substantially expands the scope of landscape genetics, allowing researchers to examine consequential questions in the study of biological diversification, including whether intra-specific genetic structure is correlated with geographic distance or ecological differences, and which ecological axes are associated with genetic differentiation in each of these species.
Training goals include gaining expertise in evolutionary ecology and in modeling and analyzing ecological traits and the interactions between organisms and their environments. Broader impacts include contributing materials to museum displays, including a planned exhibit on Anolis lizards at the Harvard Museum of Natural History, and involving underrepresented undergraduate students in several of the diverse components of this research.
Understanding how geographic, ecological, and environmental variation shape patterns of biological variation is a central goal of ecology and evolutionary biology. One particularly compelling question in modern biology is how these variables contribute to genetic diversity in wild species. Investigating patterns of genetic diversity can tell us much about how genetic variation arises, and maintaining genetic variation is critical to conservation management efforts and the long-term persistence of threatened and endangered species. In this research, I developed a novel method for measuring how different geographic and environmental variables contribute to genetic variation in nature. Different populations of a species may have different genetic compositions, which may result from their isolation in space or from the environments in which they live. Specifically, patterns of genetic variation often reflect spatial variation in gene flow and selection, both of which are shaped by environmental and geographic heterogeneity. Gene flow often acts as a homogenizing force that limits the evolution of variation among populations. When gene flow is reduced and this constraint is removed, populations will typically experience greater genetic and morphological divergence, either through drift or selection. Thus, examining how ecological landscapes influence gene flow and selection can provide key insights into the earliest stages of biological diversification in the wild. My method uses structural equation modeling, an advanced statistical technique, to test whether genetic diversity results from the effects of geographic isolation or environmental variation. It integrates a wide variety of variables, including genetic data and geographic information systems (GIS) data layers, to understand spatial patterns of biological diversity. It then uses a series of regression and model-fitting analyses to quantify the relationships between all of those variables, or essentially, how much one depends upon the others. With collaborators from Harvard University and the University of Rochester, I explored the utility of this method by examining 17 widespread species of Greater Antillean Anolis lizards. I found that both geographic and environmental variables contributed significantly to genetic divergence (36.3% versus 17.9% of variance, respectively), and these results were surprisingly consistent across most species. Overall, this analysis provides the first comparative, quantitative evidence for the prevalence of both geographic and ecological divergence during an adaptive radiation and show that understanding the evolution of spatial genetic variation requires examining all of the ways in which ecological landscapes influence patterns of gene flow among populations. These new methods and results have important implications for how we understand the nature of biodiversity. By assessing the factors that contribute to biological diversity, we can work to maintain genetic diversity and make predictions about how land-use and climatic changes will affect biodiversity. These tools should be broadly useful to wildlife and conservation managers, as well as those who seek to assess the spatial relationships of many different variables. For ecological, evolutionary, and environmental research, this will provide a significant step forward in the quest to understand the relationships in complex systems. Furthermore, the structural equation modeling approach I developed has implications for many different disciplines. Because the method is principally aimed at disentangling the effects of many different variables, it can, in theory, be applied to any complex, multivariate system. For instance, I have recently discussed this approach with researchers from the Harvard School of Public Health, and together, we are considering applying it to a study of the epidemiology of malaria in east Africa. In some highland regions of Kenya and Tanzania, malaria has re-emerged after being absent for several decades. This could have resulted from environmental warming in these areas, which result in improved growth conditions for the vector and parasite, or from increased human mobilization in rural Africa, which could have resulted in increased importation of malaria from other regions. The structural equation modeling method I developed may be able to disentangle the effects of these two processes, which could contribute substantially to our understanding of how malaria has re-emerged in these regions and for predicting threats to other regions.
