This study addresses the problem of amino acid compositional bias in molecular phylogenetic reconstruction based on protein sequences. Compositional bias is a factor that has largely been left out of conventional phylogenetic analyses, but there is now reason to believe that it is a significant factor. It is shown that in animal mitochondria homologous genes that differ in AT/GC content code for proteins differing in amino acid content in a manner which reflects the AT and GC content of the codons. This relationship is also seen in nuclear and bacterial genes, and is significant even in the highly-conserved hsp70 and $ef 1\alpha / tu$ genes. This bias in the amino acid content of proteins can result in incorrect protein-based phylogenetic trees. A striking example is presented where common phylogenetic tools fail to recover the correct tree from animal mitochondrial protein sequences. The two taxa with the greatest compositional bias continually group together in these analyses, despite a lack of close biological relatedness. It is concluded that, while protein-based phylogenetic analyses have been favoured over analyses using biased DNA, even protein-based trees can be misleading. Current models of protein sequence evolution used for phylogenetic reconstruction assume that the amino acid composition is stationary. In order to accommodare evolution involving compositional change, directional mutation probability models are proposed and their properties described. In simulation studies with biased and non-biased branches on the same tree, the biased branches "attract" in subsequent analyses, paralleling results about using real mitochondrial sequences. Improved methodology is proposed which can use both stationary and directional models in phylogenetic reconstruction, and which can use different models on different branches of the same phylogenetic tree.
| Identifer | oai:union.ndltd.org:uottawa.ca/oai:ruor.uottawa.ca:10393/9779 |
| Date | January 1997 |
| Creators | Foster, Peter G. |
| Contributors | Hickey, D., |
| Publisher | University of Ottawa (Canada) |
| Source Sets | Université d’Ottawa |
| Detected Language | English |
| Type | Thesis |
| Format | 78 p. |
Page generated in 0.0018 seconds