Research Interests

Are synonymous mutations under selection and if so why?  Are genomes randomly arranged assemblages of genes or is gene order non-random? Why are most genes apparently “redundant”?  How can we account for variation between genes in their rates of evolution?  Is the way a protein evolves solely a function of selection on the mature peptide or does the manner in which a gene and an mRNA is processed affect the final protein? My research approaches these and related questions employing mathematical, bioinformatical, systems biological and experimental tools.
Many features of genetic systems appear counter-intuitive.  Why, for example, do many single celled organisms have just two mating types and inherit their organelles from just one parent, even when both gametes are the same size? Many other issues question both the strength of selection and our understanding of how genes and genomes function.  I wish to understand whether genes, genomes and genetic systems are shaped by selection (and if so why) or whether they are neutrally evolving traits. I have shown, for example, how the genetic code is structured in a manner that ameliorates the impact of mistranslation. More recently I have been concerned with the evolution of redundancy, of gene content, of selection on synonymous mutations (and codon usage bias) and gene order evolution. A decade ago it was considered that in mammals both gene order and synonymous mutations were neutrally evolving. The research of my group has been important in questioning the received wisdom.  There is now, for example, much evidence that synonymous mutations are under selection, often owing to their effects on splicing.  Gene order evolution and mechanisms for selection on synonymous mutations are the focus of much of our current work.

The majority of my group’s work involves comparative genomics.  We are also, however, interested in a new(ish) approach to evolutionary genetics that, rather than assuming fitness parameters are known, instead
models the underlying biochemistry to derive fitness de novo.  We have employed this approach to understand why some genes have a strong knockout phenotype but most do not and why most, but not all, knockouts are recessive. It can even predict which genes an organism might have, given its ecology.  With my (very much) more mathematical collaborators (Rob Beardmore and Ivana Gudelj), we have shown how the same bottom-up approach to modelling fitness can be used to predict the outcome of evolution (in the lab), infer the generalizability of experimental evolutionary models and to understand why classical models (e.g. of co-operation) have sometimes failed to predict experimental results.

I am also the director of the GEVOteach program.  This is a research program to determine the best way to teach both genetics and evolution to school children.