Genome-level homology and phylogeny: Case studies using Vibrionaceae and Shewanellaceae (Gammaproteobacteria).

摘要

In this dissertation, genome--wide homologies are detected across complete genome sequences in two families of Gammaproteobacteria: Shewanellaceae and Vibrionaceae (Alteromonadales and Vibrionales, respectively). For Shewanellaceae (chapter one), 19 ingroup genomes are included in addition to 3 outgroup taxa. 243 locally collinear blocks (LCBs), multi--gene fragments within which there has been no rearrangement, were found representing approximately 30% of the genomes. A phylogenetic hypothesis was generated for the concatenated alignment using maximum parsimony (MP) and maximum likelihood (ML) that never matched the topologies of individual LCB phylogenetic analyses. The tree produced by concatenation could only be matched by randomly selected subsets (smaller than some LCBs) of nucleotides from the concatenated alignment. A further investigation of within--LCB gene trees continued to show this pattern of dramatic incongruence (chapter two) in that adjacent genes were never congruent with one another or with their LCB of origination in any of the 55 investigated cases.;Based on these results and on longstanding philosophical controversies over the nature of genes, it is argued that genes as units cannot be considered homologous, that only the nucleotide level is appropriate for the homology relationship (chapter three). Genes, instead, are best described as non--material, open--ended processual systems that instantiate homeostatic property cluster natural kinds. For Vibrionaceae, a more complex case due to the existence of two chromosomes, two datasets are considered (chapter four), one, as in chapter one, in which those genomes that are complete in terms of contig arrangement (19 taxa) were included and a second dataset (44 taxa) in which those genomes that still contain more than two contigs are considered. Three new complete genome sequences are presented for Photobacterium gazogenes, Photobacterium logei, and Vibrio costicola. Incongruence among data partitions obtains just as in chapter one. The 19-taxon and 44-taxon trees based on the large chromosome are largely congruent while those based on the small chromosome are not. A novel placement of three enigmatic species of Vibrionaceae, two of which are newly sequenced here (P. gazogenes and V. costicola), calls into question the current delimitation of the genus Photobacterium. One of the results of chapter four, that the second, smaller Vibrionaceae chromosome is difficult to homologize across all species is further explored in chapter five. It is shown that there is no core genome for the small chromosome across Vibrionaceae and that primary homologies are distinct from gene content comparisons, reinforcing the conclusions of chapter three, that genes are not appropriate units of homology.