In this thesis I present a novel tree based ortholog inference method, that does not require a priori knowledge of the species tree, and is thus applicable to analyse data from species of uncertain phylogenetic position. I apply this ortholog prediction method, within a fully automated phylogenetic analysis pipeline, to analyse Nemato- morph and Xenoturbella EST data in order to establish their positions within the animal phylogeny. I further extended the ortholog prediction method to a software I named POPE (Phylogeny, Ortholog and Paralog Extractor), which aids phyloge- netic analysis by providing with an integrated environment where the underlying data can be sorted, viewed and modified. In order to highlight the difficulties in obtaining a reliable set of orthologs I compare the outcome of two widely applied ortholog prediction methods, Ensembl and Inparanoid, and show how POPE can be used to filter out the cases where the two methods agree and further clarify the examples where they disagree. For the final part of this thesis I developed a second algorithm to detect lineage spe- cific substitutions within 9 Bilaterian species, where the main aim was to identify deuterostome and protostome specific substitutions of potential functional impor- tance. Furthermore, I establish the frequency of two deutero- and protostome sub- stitutions occurring within the same gene, and estimate whether they are potentially interacting in 3-dimensional space by mapping them onto known protein structure. Some of the 9 Bilaterian proteomes, required substantial pre-processing before be- ing suitable for further analysis, which in turn prompted the development of a third algorithm, which identifies and merges partially overlapping protein fragments.