The overarching aim of my research is to understand the genetic basis of phenotypic variation within species. Genetic factors influence many phenotypes of biological and medical interest, including susceptibility to most human diseases. A central goal of biomedical research is to identify these factors, as a crucial step in improving diagnosis, prevention, and treatment. The complex genetic basis of most traits involves multiple genes that interact with each other and the environment. My lab conducts experiments in model organisms (currently, the yeast Saccharomyces cerevisiae and the nematode worm Caenorhabditis elegans), as well as computational analyses, aimed at understanding how changes at the level of DNA are shaped by molecular and evolutionary forces and how these changes lead to all the observable differences among individuals within a species.
Yeast Genetics and Genomics
Genetic mapping of complex traits. Our lab pioneered research on genetics of global gene expression with a study in yeast that characterized the different classes of regulatory variation and showed that inheritance of transcript levels is typically complex. We have identified and validated specific polymorphisms responsible for widespread expression changes and showed that trans-acting regulatory loci are not predominantly transcription factors. We have elucidated general principles of the genetic architecture of complex quantitative traits. We have developed new statistical and computational methods for identification of genetic interactions, showed that such interactions play a role in the genetics of 50 percent of expression phenotypes, and validated one such interaction at the level of the molecular polymorphisms. We have used classic cis-trans tests to determine the contribution of cis-acting polymorphisms to variation in transcript abundance.
More recently, to investigate the interactions between genes and environment, we have grown strains in two different conditions—fermentation and respiration, measured global gene expression, and characterized the prevalence and strength of interactions between genetic loci and environment, as well as identifying specific examples of such interactions. We have also used genetics to understand other quantitative phenotypes in yeast, including telomere length and sensitivity to small molecules.
We are expanding our studies of the genetics of sensitivity and resistance to drugs and other chemical compounds. Currently, we are focused on improving prediction of cellular response to these compounds based on gene expression, genotype information, and a combination of the two. The ability to do so would provide insights into selecting treatment based on an individual's genotype at the relevant loci or a tumor's molecular signature. Our approach also allows us to investigate the precise genetic basis of strain-specific phenotypic effects of prion-like elements in yeast. We have engineered the two parent strains to carry mutations in the translation termination factor Sup35, which is the protein whose misfolded forms correspond to the PSI+ prion state, and we have shown that these mutations can confer growth advantages or defects in response to different growth conditions in a manner that depends on the genetic background.
To map and identify the polymorphisms between strains that are responsible for the differential responses, we are now engineering the segregants and testing them in different conditions. In collaboration with Susan Lindquist (HHMI, Massachusetts Institute of Technology), whose lab pioneered work in this area, we are investigating the PSI+ state in diverse yeast strains and investigating genetic differences in other phenotypes associated with protein misfolding. We are also extending our work on global gene expression levels to measurements of protein abundance and post-translational modification, as well as to assays of cellular metabolism.
Yeast population genomics. Studies of genetic variation are shifting from those carried out in families (or crosses) to those carried out in populations. In human genetics, the main driver of this shift is the greater power of population studies to detect common alleles with modest effects on disease susceptibility. This alone is reason enough to model such studies in simpler systems. Furthermore, studies of variation in natural populations provide insight into the evolution and distribution of polymorphisms with phenotypic effects that cannot be obtained from family studies or crosses. Thus, we are expanding our studies of variation in yeast from a cross between two strains to a large collection of diverse strains.
We are currently using the tiling array system we developed for genome-wide polymorphism surveys to characterize the genomic diversity of yeast strains sampled from all over the world and from different ecological niches (beer, bread, vineyards, immunocompromised individuals, various fermentations, and environmental isolates). We are analyzing the evolutionary history of the S. cerevisiae species, the patterns of linkage disequilibrium, and the forces that influence sequence diversity, including selection. We plan to extensively phenotype this collection of strains, measuring global gene expression and other traits. To capture the diversity of ecological niches in which these strains naturally find themselves, we are developing assays of growth under a broad range of conditions. To understand the molecular basis of adaptive evolution, we aim to connect sequence variation across the species with variation in gene expression, ecological niche, and growth in different conditions. This work includes investigation of statistical and computational methods for association studies in structured populations that will prove useful for studies in humans and other species. (Our studies in yeast and statistical and computational methods development receive support from the National Institutes of Health and the James S. McDonnell Foundation.)
Complex Trait Genetics in C. elegans
Choice of C. elegans as a model. C. elegans is a well-established model system for studies of many aspects of metazoan biology, including development, behavior and its neural basis, and aging. As such, it provides a rich set of phenotypes that are relevant to higher organisms. C. elegans occupies an important middle ground between the detailed studies of genetics of gene expression that my group has carried out in yeast and studies in more complex organisms, principally mammals. Short generation time, small size, simple genetics, and powerful genomic resources allow large studies and facilitate high-resolution mapping and identification of specific nucleotide changes responsible for phenotypic effects to an extent difficult to realize in mammalian systems.
At the same time, C. elegans is a metazoan, and thus allows studies of more complex gene regulation and biology not possible in yeast. For example, regulation by microRNAs is important in C. elegans but absent in yeast. Gene expression can be studied at different developmental stages and in different cell types. Other important organismal phenotypes can be studied in parallel, allowing connections to be made between genetic variation in global gene expression and variation in morphology, development, physiology, and behavior. Many natural isolates have been identified and found to vary in a large number of traits, including gene expression, brood size, male mating success, rate of spontaneous male production, susceptibility to RNA interference (RNAi), resistance to pathogenic bacteria, and a variety of behaviors.
Genetic resource construction. We constructed a set of recombinant inbred lines (RILs) by crossing the natural isolate N2 (Bristol), most commonly used in lab studies, to the most divergent known isolate CB4856 (Hawaii), which differs from Bristol at roughly 1 nucleotide per thousand. To achieve high genetic resolution, a large population of worms was intermated for 10 generations to maximize recombination breakpoints, and then individual worms from the F10 generation were inbred to obtain a total of 239 homozygous RILs, each representing a different mosaic of the two parent genomes. We also assembled a diverse collection of wild isolates to facilitate population studies. We genotyped the 239 RILs and 127 wild isolates at 1,536 uniformly spaced single-nucleotide polymorphisms (SNPs), using Illumina's GoldenGate assay, and built a high-resolution genetic map of the RILs. The RILs provide a powerful resource for genetic mapping of the many traits that vary between Bristol and Hawaii and segregate in these lines. We plan to build additional RILs for these and other parent strains.
A common genetic incompatibility in C. elegans. During the construction of the RILs we discovered a previously unnoticed genetic incompatibility among C. elegans wild isolates. This incompatibility causes embryonic lethality in the F2 generation of crosses between Bristol and Hawaii. Two linked loci are involved: the Bristol allele of one locus acts via paternal effect to induce lethality of embryos homozygous for the Hawaii allele of a second, zygotically acting locus. We have cloned the zygotically acting locus by transgenic complementation. The Bristol and Hawaii haplotypes of the incompatibility locus display a level of polymorphism far above the genome-wide average. We have found that both Bristol-like and Hawaii-like strains are common and co-occur in natural populations. The high level of polymorphism at this locus indicates that both haplotypes are ancient, yet the fact that the Hawaii-like haplotype is deleterious when crossed to a Bristol-like strain suggests that natural selection should rapidly eliminate it. Thus, the long-term maintenance of both haplotypes must be due to balancing selection. We are working to characterize the molecular mechanisms involved in the incompatibility. We also seek to understand the specific evolutionary forces that are maintaining the incompatibility.
Transcriptional variation. We plan to study the genetic basis of differences in gene expression in C. elegans. In preliminary experiments, we have shown that more than 3,000 genes are differentially expressed between Bristol and Hawaii worms at the young adult stage. We have initiated expression profiling of the RILs at this stage, and we plan to analyze both the general quantitative genetics of the thousands of expression traits and the molecular genetic basis of expression differences in a manner analogous to our studies in yeast. We will extend expression profiling to other developmental stages, as well as other environmental conditions. We will also investigate approaches for measuring expression in specific tissues.
Phenotypic variation. The genotyped RILs are an important resource for the study of many types of phenotypic variation in C. elegans. We recently identified a transposon insertion into a novel gene encoding a mucin as the molecular basis of a classic dimorphism in male mating. Males from many natural isolates (including Hawaii) deposit a copulatory plug after mating, whereas males from other natural isolates, including Bristol, do not deposit plugs. Although C. elegans descends from an obligate-outcrossing, male-female ancestor, it occurs primarily as self-fertilizing hermaphrodites. The reduced selection on male-male competition associated with the origin of hermaphroditism may have permitted the global spread of a loss-of-function mutation with no evident effects beyond the loss of male mate guarding.
We are investigating the genetic basis of a number of other phenotypic differences in physiology and behavior. These include behavioral differences such as attraction to and avoidance of various odorants, related memory and learning behaviors, locomotion, mating behaviors, and male-specific behaviors. We are developing automated microscopy and digital image analysis systems for quantitative measurements of behavior and movement. We are also investigating differences in susceptibility to microbial pathogens, in stress resistance, and in longevity. We have established collaborations with several worm labs that specialize in studies of C. elegans biology and behavior. Our current collaborators include Cornelia Bargmann (HHMI, Rockefeller University), Patrick Phillips (University of Oregon), and Coleen Murphy (Princeton University). We plan to profile the RILs and wild isolates for a broad range of molecular, behavioral, developmental, morphological, and physiological traits. Collection of multiple classes of phenotypes, both molecular and organismal, in the same study samples will allow us to draw connections among the multiple levels of phenotypic information. Our goal is to understand how changes at the DNA sequence level are translated into changes at the level of organismal phenotypes, through changes at the molecular and cellular levels. (Our studies of transcriptional and behavioral phenotypes in C. elegans receive support from the National Institutes of Health.)