In this thesis, three distinct studies were carried out to investigate various aspects pertaining to the properties and applicability of phylogenomic data used in a supertree context. While the availability of genomic scale data is rapidly diminishing as a problem in the field of phylogenomics, there is now a greater need for an appropriate means of analysing such data. Supertrees have emerged as a useful approach in handling large data sets and have been shown to work extremely well in a phylogenomic context (e.g. Creevey et al., 2004, Fitzpatrick et al., 2006, Pisani et al., 2007). While supertree studies do generally sample significantly more genomic data than their supermatrix counterpart, much of the genome, which has evolved in the light of gene duplication, is not considered in this method. Further to this, typically, in the supertree approach complete genomic data is exclusively used, which can result in a very limited taxon sampling compared to alternative approaches that use expressed sequence tag (EST) data. Here, in attempt to address these shortcomings, the viability of integrating genes with a history of duplication in the supertree approach, as well as the extent to which a combined data set of complete and partial genomes (ESTs) can be used to increase taxon sampling in this context, is investigated. Additionally, in this thesis, the effect of input tree shape biases is assessed. It has been shown previously that some commonly used supertree methods are biased with respect to the tree shape they produce (Wilkinson et al., 2005). However, since some supertree methods (e.g. matrix representation with parsimony; MRP) have an inherent phylogenetic component, the observed shape predispositions of these supertree methods may be attributed to such methodological elements. As such, here the effective shape bias of various phylogenetic methods is assessed using a phylogenomic data set.