Behold the power of Conjugative Transfer
Ecological and Evolutionary processes in the long-term evolution experiment
Working with Jeff Barrick and Austin Meyer, I collected data on how allele frequencies change over time in the Ara-1 lineage of the long-term evolution experiment. These data are inconsistent with simple models of asexual evolution that have been fitted to these data. This led us to hypothesize that transient frequency-dependent interactions evolve on short time-scales, resulting in the observed deviation from theory. I confirmed the existence of these interactions, and probably figured out their genetic basis.
Comparative Genomics With The Long-Term Evolution Experiment
The genes under positive selection in the long-term E. coli evolution experiment (LTEE) tend to be fairly conserved in nature. These loci tend to be less polymorphic, compared to other genes in E. coli, and less diverged, compared to other genes conserved since the split between Salmonella typhimurium and Escherichia coli. Further, the "core genome" of Escherichia coli -- the set of all genes found in all Escherichia coli -- is under strong positive selection in the LTEE. At first, this might seem to be a surprising result. However, core genes also tend to be functionally important genes; knockouts of these loci tend to have strong growth defects, for instance. In short, genes important to E. coli's physiology are under positive selection in the LTEE, and these genes evolve fairly slowly in nature, probably because they are important in defining E. coli's niche in the wild. Life in a flask, even at biological standard conditions, is a strange place for E. coli to find itself over evolutionary time scales.
Mutation rates, Synonymous Diversity, and Effective Population Size
Levels of synonymous genetic diversity accumulate proportionally the product of effective coalescent population size and mutation rate. This measure is not uniform over E. coli's genome. Some workers have proposed that the mutation rate varies fairly dramatically over the genome. Alternatively, this result could be due to differences in the effective coalescent population size across loci in E. coli. Linkage disequilibrium does not seem to implicate selection as the mechanism. My pet hypothesis is that different loci in E. coli can have different census population sizes, depending on their history of horizontal and vertical transmission within microbial communities. Bacterial genomes contain information about species phylogeny (vertical descent) , as well as the history of transmission of alleles across species in communities (horizontal transmission.)
The long-term evolution experiment is a powerful model system both for dissecting evolutionary and ecological processes, and for testing theory. But what are the limitations of this model for understanding how microbial communities evolve in the real world? In my thinking, the most obvious difference between dynamics in the experiment, versus the real world, is the lack of recombination, horizontal gene transfer, and of viruses that can act as vectors for strange genetic material. I’m interested in comparing the dynamics of genome evolution between wild E. coli and the long-term experiment to explore these issues. I'm also interested in using model communities of digital organisms to study these issues without having to deal with sampling problems. I am particularly interested in how multi-level selection shapes genome organization. My working hypothesis is that bacteria are porous organisms, akin to communities of genes that work together, often susceptible to cheating as well as cooperation on the gene level. This view of evolution as operating on multiple, recursive levels is deeply appealing to me, intellectually. I want to explore the plausibility of some predictions of this model, such as 1) some genes should have larger effective population sizes than the organism they live in; 2) different genes might have different strategies for survival in a community; 3) many genes in bacterial genomes might not be adaptive—and might even be maladaptive as genetic parasites; 4) genomes and genetic elements in general might be intrinsically modular in terms of how cooperative their dynamics are; and 5) The network of genetic interactions between alleles should be structured by genes evolving to be resistant to replacement by alleles from the outside, while simultaneously trying to be able to replace alleles in other genetic backgrounds. I think these ideas could have deep implications for our understanding of sex and speciation.