 |
 |
 |
 |
 |
 |
 Research Overview Previous Research Behavioral Ecology Statistical Genetics Population Genetics Current Research Canalization and Genetic Networks Evolutionary Ecology of Virulence Evolution The Micro-Evolution of Regulatory Control The Evolution of Gene Families Research Overview 
In my research I use mathematical and computational tools to solve problems in genetics and evolutionary biology. My background includes formal training in both biology and pure and applied mathematics. This dual-path training has given me a comprehensive set of techniques that allows me to find and use the best method for solving new biological problems, crafting elegant solutions that encompass the appropriate level of complexity.
Building on my research experience in behavioral ecology, statistical genetics, and population genetics, I am currently examining the evolution of genetic networks, genome structure, and developmental regulation. This focus has developed during the past several years of my association with the University of Oregon, where I have been an active member of the NSF-funded IGERT program in Evolution, Development, and Genomics. My current research is funded by an NIH NRSA fellowship and promises to be an area of increasing interest. Previous Research Traditional models of male display assume that display traits are made once, last forever, and have a fixed cost. I have created game theoretic models of male display that incorporates the timing of signaling costs and the effects of aging. This work shows that signaling is generally more ``honest'' among older males, which is a novel explanation for the observed female preference for older mates (Figure 1). Because the amount of future reproduction available decreases as individuals age, the optimal signaling strategy shifts to higher levels of signaling for all males. However, the relationship between quality and signal changes as males age so that a tighter relationship between male quality and signal exists in older age groups. |
I have worked with colleagues in molecular genetics to understand the functional relationship between gene products at the synapse of C. elegans. I designed a maximum likelihood method to test for genetic interactions between partially recessive mutations at known loci. The estimates of the genetic interactions were obtained from an analytical model, but a computational approach was required to define confidence intervals and conduct hypothesis testing. By applying this method we were able to both verify known physical interactions (validating the method), and infer novel interactions between gene products.
My research in population genetics has included studies of sexual selection, sex allocation, genetic dynamics in stochastic environments, and virulence evolution. A major theme of my research is concerned with the interaction between sexual selection and adaptation in both space and time. I have shown that the mating system can have a large influence on whether populations survive climate change and affect the rate of adaptation to novel conditions. Further, sexual selection can enhance the ability of a species to achieve local adaptation and maintain genetic variation along a spatial gradient. This creates a feedback cycle that strengthens selection for female mating preferences and local adaptation simultaneously. By focusing on the dynamics of adaptation and mating system evolution, I have shown that speciation is not necessarily the most likely response to spatial variation, but rather that sexual selection can maintain both local adaptation and gene flow simultaneously.
I am generally interested in the role that stochasticity plays in adaptive evolution and the creation of neutral genetic variance. As biologists we often attempt to predict the outcome of evolution as though it were a deterministic process, despite the fact that it is inherently stochastic. A more valid view of adaptive evolution embraces its inherent uncertainty, focusing on the relative probability of observing different adaptive states. I have taken this approach by studying trait evolution both when individuals face random variance in their ability to reproduce and when whole populations face random environmental variance. For the case of individual variance, I have derived both methods to bound the fixation probability of a novel allele, as well as diffusion approximations to calculate fixation probabilities. This method represents an important improvement on previous methods, as is shown by its ability to predict simulation results (Figure 2). I have applied this method to sex allocation in finite populations and have shown that allocation to the more risky sex will be reduced, a pattern observed in plants with risky pollination syndromes.
Figure 2: The curve represents the predicted fixation probability from the diffusion approximation while the dots represent observed fixation probability from
simulations.
Canalization and Genetic Networks
One of the central themes of organismal biology is the functional integration of the organism as a whole. Changes in one part or system of an organism are likely to have cascading effects through its other systems and impact seemingly unrelated functions. These cascading effects make the organism susceptible to perturbations in the environment or in the genome. On the other hand, regulatory interactions among different systemic elements allow the organism to lessen the impact of these cascading effects through buffering, feedback, and compensation. Genetic canalization can be defined as the evolutionary reduction in the effect of perturbations on the output of a genetic network.
I have developed a general framework for studying the strength of selection on canalization. The simple result is that selection on canalization is at most equal to the fitness load minus the per gene mutation rate (L-m). For load induced by mutation alone, canalization is only favored when the number of interacting genes is large, as is the case for heat shock proteins. However, the fitness load induced by both spatial and temporal variance can be large, providing ample selection for canalization even when only a few genes are involved. This implies that ecological processes play an important role in shaping genetic interactions and genomic structure.
Evolutionary Ecology of Virulence Evolution
Virulence evolution is an intrinsically complex problem because it includes several interacting partners, hierarchical spatial structure, and selection at multiple levels. The dynamics of virulence evolution cannot be modeled with standard evolutionarily stable strategy techniques because of both within host dynamics and the effects of timing in transmission. My recent work with Troy Day has created a general framework to study virulence dynamics that includes within host population genetics. We modeled both super-infection and within host mutation using the same technique and were able to show that within host mutation generally leads to an increase in virulence.
I am expanding on this theme to consider models that involve both host population dynamics and virulence evolution. Because the selective pressure on increasing virulence depends on host population size, changes in virulence caused by mutation and super-infection will lead to changes in host population size. This can lead to a kind of feedback loop, not unlike mutational meltdown, that causes an open ended increase in virulence. This theory could explain the existence of highly virulent pathogens that become locally extinct.
The Micro-Evolution of Regulatory Control
The regulatory control of genes involved in development has recently received increased attention due to experimental advances. One of the striking features is that positive regulation occurs at a broad level while inhibitory regulation is more important in determining where genes are expressed. I am currently developing models that include both population genetics and developmental dynamics to investigate the micro-evolution of regulatory control during development. My preliminary results suggest that positive regulation of spatial expression can only evolve by major changes, while inhibition control can evolve by small steps. This suggests that the present patterns of genetic control evolved because of historical constraints rather than due to the system level properties of control structures.
The Evolution of Gene Families
Most evolutionary models of gene duplication focus on events that occur after a duplication has become fixed in the population, and largely ignore the role of natural selection in setting up the conditions that are favorable to gene duplication. This has created a paradox where gene duplicates and gene families evolve more commonly than the neutral approach predicts. I have derived a population genetic model that include selection before, during, and after duplication to examine the role that natural selection plays in the evolution of gene families. My main result is that the divergence and specialization that creates gene families usually begins when the gene is present as a single copy. This divergence creates strong selection for duplication, in most cases leading to a smooth evolutionary birth of a gene family before duplication (Figure 3).
My results suggest that the creation of gene families is not caused by relaxed selection following duplication events, but rather by selection for genetic specialization within an organism. When divergence occurs at a single locus, heterozygote advantage is created. Once heterozygote advantage has evolved, there is selection for gene duplication through a form of canalization. Because natural selection plays a prominent role in this theory it can help to explain why so many gene duplicates are maintained over evolutionary time.
Figure 3: In the green region, selection for genetic
specialization causes allelic divergence at a single locus. In the yellow
region, divergence and duplication occur simultaneously. Only in the small
red region does divergence occur after a gene duplication is fixed
in the population.