Emily Hornett

I am interested in the population genetics and the evolutionary dynamics between species in natural populations. I have undertaken a lot of fieldwork over the years, and enjoy investigating the complexities of natural systems.

The interaction between Hypolimnas bolina and Wolbachia

Much of my work to date has been on the dynamic interaction between the butterfly Hypolimnas bolina and the endosymbiont Wolbachia. This bacteria is very common in insects, and causes a multitude of effects - many of which we are still discovering. It was initially studied due to its ability to manipulate the reproduction of its host to favour the production or fitness of females - through which it is transmitted. More recently it has been shown, along with a growing number of other influential symbionts, to also have the ability to provide pathogen or parasite resistance in the host.   In the tropical butterfly H. bolina it causes Male-killing, where the bacteria kill male offspring early in development and consequently alter the sex ratio of the brood, and if at high frequency, the population. A distortion of the sex-ratio has long been theorised to induce strong selection in order to return it to parity (eg Fisher 1930). Early on in my work we discovered that in some populations of the butterfly this bacteria did not kill males yet still infected the population at high frequency. We demonstrated for the first time that a dominant host suppressor of the activity of male-killing has evolved in some localities. In addition, we showed that the bacterium could persist despite losing its male-killing ability by having a second 'hidden' phenotype called Cytoplasmic Incompatibility (a common manipulation of Wolbachia in many insects where uninfected females produce no or few offspring upon mating with an infected male).  Before this finding it was thought generally the case that these endosymbionts have only one trick in its bag (one major phenotype)....now known through multiple studies to not be the case.

Following this I moved onto the genetics of the suppression, locating the suppressor to within a 4cM region of the genome, and investigating the effect of the rapid spread of the suppressor which we were lucky to observe in 'real-time', on the rest of the host genome. In 2001 we observed a 99% frequency Wolbachia infection in H. bolina in the islands of Samoa in the South Pacific. This caused a sex-ratio bias of 100 females to every  male. We also estimated, from historical records, that this extreme sex-ratio bias has lasted for over 100 years, which in itself was remarkable (Dyson et al 2004). In 2006 we received word that males were seen in increasing numbers on the main island, and after  fieldwork there we demonstrated that a suppressor, potentially the same as the one found in SE Asia, has invaded and spread through the population at a rapid rate. When looking at loci linked to the suppressor I observed that almost the entire length of the chromosome on which the suppressor was located had hitchhiked along with the suppressor loci and swept through the population. This is likely one of the largest natural selective sweeps observed to date, and evidences the widely held view that parasite/host interactions, particularly those that distort the sex-ratio, have the potential to produce very strong natural selection.

Bioinformatics study of the application of RNA-seq in non-model species

I am interested in how well RNA-Seq and indeed any of the latest sequencing technologies and applications perform when your study system does not have any genomic resources, i.e. its not a model species. In this study I, with Christopher Wheat at Helsinki/Stockholm University, used the resources and data available for H. sapiens to compare the direct mapping of sequencing reads to predicted genes from the genome, to de novo transcriptomes assembled from RNA-seq data. Further to this, and pertinent to studies of non-model species, we then investigated gene coverage and expression analysis using increasingly divergent (from H. sapiens in this case) species with genomic resources to group contigs in the de novo transcriptome by unique genes...a necessary step in many studies. We estimated the amount of error and bias encountered by using this approach and asked how far in evolutionary divergence terms can we go in using proxy reference species to annotate expression data. This work is currently in review at BMC Genomics.

Comparative genomics

More recently I have broadened my research to multiple species of butterfly. My current projects include comparative genomics of 12 butterfly species, in particular estimating evolutionary rates of various targeted genes, and investigations into the maintenance of polymorphisms in natural populations. I have also continued my interest of species interactions and am studying the bacterial infections of these species using a suite of bioinformatics tools to data-mine Illumina datasets.