This week’s BEACON Researchers at Work blog post is by University of Idaho postdoc Matthieu Delcourt.
The evolution of complex traits is one of the major enigmas in evolutionary biology. While we know a lot about phenotypic variation within populations, much remains to be explored about the genetic causes and consequences of trait complexity. We know that the genome is composed of many genes that are located on same or different chromosomes and that their proximity influences their respective evolutionary histories across many generations. We also know that these same genes can interact together by repressing, activating or regulating protein expression. Taken together, these types of interactions between genes can affect phenotypes dramatically and this may also vary between individuals. Quantitative traits such as body size, life history traits or stress tolerance are good examples of traits that often show a great amount of variation within a population, and are also good examples of how genetic interactions can therefore influence the process of adaptation. Quantitative traits depend on the action of many interacting genes with small effects that often generate genetic correlations between multiple phenotypes. Those interactions can happen either at the chromosomal level or the cellular level and can either constrain or increase the nature of genetic variation that is relevant to selection.
In Dr. Paul Hohenlohe’s lab at the University of Idaho, we explore the effects of gene network structure on the adaptation to stress tolerance in the budding yeast (Saccharomyces cerevisiae). We are interested in linking phenotypic to genomic patterns of variation within the context of quantitative genetics. These are very different approaches that gather information at various levels and that are often difficult to combine in a series of experiments. We have started a project looking in particular at the effect of size and degree of independence of gene networks involved in stress tolerance in salt, glycerol, and copper sulfate. We take advantage of the yeast as a microorganism to perform experimental evolution, phenotypic and genomic analyses in a polymorphic population that we created in our laboratory. This population has shown a large amount of phenotypic and genetic variation in stress response, meaning that the underlying genetic architecture differs between individual cell lines within that population. Interestingly, the level of connectivity between networks seemed to directly affect the genetic correlation between traits. Take the example of two osmotic stressors (salt and glycerol): the cellular response will involve very similar sets of genes (that are tightly connected) to equilibrate the osmotic pressure inside the cells in either case. On the other hand, when the gene networks involved are very distinct, for example when one network is involved in the cellular response to an osmotic stress (salt) and the other to an oxidative stress (copper sulfate), then very different sets of genes are regulated. We found that in the first case (see figure above), the stress responses were genetically correlated, meaning that an individual cell that had the machinery to deal with salt could also deal with glycerol very well. In the second case, there was no genetic correlation between the stress responses, meaning that these network were to some extent free to evolve independently genetically. One can wonder how this fits into the broader picture of trait complexity. This is actually a very exciting result that extends previous finding on multivariate phenotypic evolution to the level of gene-gene interaction. Now that our field gathers lots and lots of genetic data, we need to make sense of the effects of all of these genes and their interactions altogether on the phenotypes. Can we predict traits in individuals given we know something about their genotypes? Many human diseases, genetic syndromes or psychological disorders indeed depend on one, two or more gene networks that are inter-connected and these are often complex traits that vary tremendously within and among populations.
We are currently exploring where in the genome this variation in stress response originates and how many mutations are responsible for this variation. Do mutations tend to hit at the center or the periphery of a gene network? How does the degree of independence between gene networks constrain the evolution of stress response in yeast? These are some of the questions that we ask in our lab and that we hope will encourage research at the interface between phenotypic, quantitative genetic and genomic analyses.
For more information about Matthieu’s work, you can contact him at mdelcourt at uidaho dot edu.