Dr. Leanne Haggerty
Leanne obtained a B.Sc. degree in Genetics and Bioinformatics at the The National University of Ireland, Maynooth. As an undergraduate she undertook two summer placements in a bioinformatics laboratory, focusing on the construction of the phylogeny of closely related bacteria. Her undergraduate thesis focused on the systematic bias that appeared when estimating phylogenies using Maximum Likelihood.
She started her PhD in the bioinformatics laboratory, under the supervision of Professor James McInerney, in 2009, funded by Science Foundation Ireland. Completed in October 2012, the title of her PhD thesis is “Exploring the Non-vertical Component of Bacterial Evolution Using Homology Network Structures”.
She currently holds a Post-doctoral research position in the bioinformatics laboratory where she is involved in studies to discover the type of evolutionary relationships that have been overlooked in previous phylogenetic studies. She is also currently developing GCUA2.0, an updated version of the GCUA software, written in Python and including additional codon usage indices.
PhD Research Focus:
The beginning of my PhD was focused on network representation of homologous relationships between members of the Proteobacteria. In reality genes are inherited vertically but can also be acquired through horizontal transfer (HGT), which does not adhere to the conventional understanding of speciation events and so is problematic when building phylogenies.
In bacteria barriers to HGT are very low, it has been estimated that 81 ± 15% of all genes in prokaryotic genomes have undergone HGT at some point in their evolutionary history. Often the fate of these genes or parts of genes is to be integrated into a recipient genome by recombination. As a result of homologous recombination, closely related strains become more diverse and distantly related species gain similarity. Since it is not limited to within-genus it becomes increasingly difficult to segregate strains based on their shared gene pools. By assessing levels of sequence similarity we can show which genes have the least divergence i.e. those that are still recombining. High levels of sequence similarity between two genomes that might otherwise show moderate levels of divergence might be explained by HGT.
In biology, networks have become popular when describing prokaryotic evolution as a replacement for the less informative bifurcating tree. The tree representation was limited to describing evolutionary relationships based on vertical inheritance, whereas a net-like diagram would encompass the horizontal acquisition of genes. On these “gene-sharing” networks, the economic agents (nodes) are genes or genomes. An edge between two genes represents a homologous relationship whereas an edge between two genomes represents the number of genes they have in common. Thus the network provides a visual representation of all gene relationships, whether acquired through vertical inheritance or horizontal transfer.
By developing a pipeline to build genome and gene networks form BLAST hits, I was able to create an all-encompassing view of a group of closely related bacteria and thus analyse the extent of gene sharing amongst said bacteria.
A network of genes from enteric bacteria with homology across 100% of the sequence
Homology networks reveal a lot about gene relationships that cannot be communicated by a tree. With this in mind I worked on developing an algorithm, based on network properties, to create a dataset enriched in fusion genes.
A fusion gene is the result of an event whereby two previously separate genes are joined to encode a single, usually multifunctional, protein. Labeled by many as ‘Rosetta stone’ proteins, they have proved to be key in finding potential protein-protein interactions and metabolic or regulatory networks. With the accumulation of genome sequence data, it has become obvious that domain, gene and indeed perhaps genome fusion is frequent in nature, though it is still not clear how often fusion events occur and what the main drivers of fusion are likely to be. Often the function of the separate genes is not detected until a corresponding fusion of these components is discovered in another genome. Current fusion detection algorithms rely on non-overlapping, side-by-side BLAST matches of two genes from a reference genome to a single open reading frame (ORF) in a target genome. This approach is cumbersome and difficult to implement on a large-scale. I propose a network-based approach to finding fusion events.
McInerney, J.O., Cummins, C.A. and Haggerty, L.S. (2012) Goods-thinking versus tree-thinking: finding a place for mobile genetic elements.
Mobile Genetic Elements 1, 4. [advance access].
Haggerty, L.S., Martin, F.J., Fitzpatrick, D.A. and McInerney, J.O. (2009) Gene and Genome Trees Conflict at Many Levels.
Philosophical Transactions of the Royal Society of London: B Series. [link]
Mechanisms of Protein Evolution, 2011
Title: A Network-Based Approach to Finding Fusion Genes
7th Annual Rocky Mountain Bioinformatics Conference, 2009
Title: Gene and Genome Trees Conflict at Many Levels
9th Annual Rocky Mountain Bioinformatics Conference, 2011
A Network-Based Approach to Finding Fusion Genes
A Network-Based Approach to Finding Fusion Genes
Young Systematists forum. London, 2010
Gene-sharing networks define species groups in prokaryotes
Society for Molecular Biology and Evolution Meeting. Barcelona 2008
A Supertree Analysis of 27 Complete YESS Group Genomes