Phylogenetic analyses for developmental evolution.

摘要

Accurate gene phylogenies are essential for reconstructing the evolution of gene expression patterns, a central component of the study of the evolution of development. Current practice is rife with easily corrected errors of implementation. Among these are the neglect of tree reconstruction methods, relying on BLASTP scores alone; the use of gap-filled computer generated alignments; the lack of use of bootstrapping to evaluate the reproducibility of particular nodes; and the acceptance of bootstrap percentages well below 50% as supporting a particular gene family assignment when bootstrapping is used. These errors are easily corrected, involving less than a day's work even for those inexperienced with phylogenetic tree reconstruction.; I present here an examination of bootstrapping methods and four case studies of developmental genes for which more extensive methods than are typically employed are used. First, phylogenetic analysis of an unusually short stretch of protein sequence can give a reliable classification of plant myosin genes, independently supported by the presence/absence of an intron. Second, the phylogenetic analysis by multiple methods at multiple hierarchical levels combined with analysis of the developmental expression of AmphiEomes/Tbr/Tbx21 are presented. Third, the question of early vertebrate whole-genome duplications is addressed, including numerous gene trees constructed to investigate this issue. Fourth, the reconstruction of the genomic structures of the amphioxus and vertebrate Tbx2/3 and Tbx4/5 loci and implications for the evolution of gene regulation are discussed. Finally, I test the relative performance of each of the three tree reconstruction methods.; In the first three cases, I have made the simple addition of using multiple tree reconstruction methods. The use of multiple methods reveals aberrant results due to the misapplication of a single method. The tendencies toward specific types of errors for each of the methods are tested and discussed in the sixth chapter. A further improvement on standard practice was shown by the fourth example in which non-parametric tests of various hypotheses were reported. These tests returned significant results concerning the relative placement of protostome and deuterostome genes, a question commonly addressed in comparative developmental studies reporting the discovery of putative orthologs. These "new" (to developmental biologists) methods detect errors in previously established orthology claims and if implemented would place studies of conserved and novel gene expression patterns on a firmer scientific footing.