February 10 2011

10:00 LSB 2320

This seminar is sponsored by The Departments of Ecology & Evolutionary Biology and Microbiology, Immunology, & Molecular Genetics

Ruth Hershberg
Department of Biology, Stanford University

Disentangling the Determinants of Genome Evolution


Mutation and natural selection, together with stochastic processes such as genetic drift,
determine the patterns of variation we observe between organisms. Both selection and mutation introduce biases to the patterns of variation observed in the genome sequences of different organisms. Selection introduces such biases by favoring certain variants over others based on fitness, which depends on function. Mutation introduces biases because certain types of mutations occur more frequently than others. Such biases in mutation can change between organisms and even within organisms in response to changes in environment and stress or in different genomic regions. The biases of mutation introduce non-random patterns to variation beyond those introduced by selection. This makes it very difficult to separate the effects of selection from those of mutation and study them in isolation. My approach to disentangling selection and mutation is to take advantage of the recent vast increase in sequence data availability to focus on evolutionary scenarios in which selection acts with reduced efficacy. In my talk I will describe how I demonstrated that several important bacterial pathogens, such as Mycobacterium tuberculosis, Salmonella typhi, Bacillus anthracis and Yersinia pestis are subject to extremely relaxed selection. This relaxation in selection leads to phenotypic consequences, such as an accumulation of functional point mutations across their genomes and accelerated gene loss. I will then describe how I used large sequence datasets of such pathogens to probe mutational biases and demonstrate that mutation is universally AT-biased across bacteria. This finding contradicts the long held view that mutational biases are highly variable among bacteria and suggests that nucleotide content may be a selected trait.















































































































































































































































































































































































































































































































this is idtest: