Improvements in codon usage analysis for a more detailed understanding of genome content and horizontal gene transfer

codon usage;Escherichia coli;Agrobacterium tumefaciens;Borrelia burgdorferi;bioinformatics

Davis, James J.

关键词: codon usage;    Escherichia coli;    Agrobacterium tumefaciens;    Borrelia burgdorferi;    bioinformatics;   

Others  :  https://www.ideals.illinois.edu/bitstream/handle/2142/15544/Davis_James.pdf?sequence=5&isAllowed=y

美国|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

The genetic code has evolved with considerable elasticity, enabling most amino acids to be encoded by multiple synonymous codons.Genes can vary in their utilization of synonymous codons, and this provides a basis of comparison for studying the compositional histories and evolution of genomes.The original goal of this dissertation work was to study the effects of horizontal gene transfer in diverse genomes; however, these efforts were quickly encumbered by limitations in the current methods of codon usage analysis.In this dissertation, we describe the limitations of these methods, and challenge the fundamental assumptions that they are based upon.In order to evaluate horizontal gene transfer (or any other source of variation within a genome) it is first necessary to define what is “typical”.Many previous studies have considered the typical codon usage of a genome to be the genome-wide average.In Chapter 2, we establish a method for calculating the modal codon usage of a genome and demonstrate that it is more resistant to the effects of aberrant genes than the average.In Chapter 3, we use the mode algorithm to study the evolution of Agrobacterium tumefaciens and Borrelia burgdorferi—two bacterial genomes that contain multiple replicons.In A. tumefaciens we discover that the two plasmids are closely related, despite being independently conjugative.By using the mode algorithm on the B. burgdorferi genome, we are able to demonstrate a higher resolution of codon usage relationships than had been previously shown—we observe a close similarity between the linear plasmid lp38 and the chromosome, and a close similarity between the members of the cp32 family of plasmids.We observe that these codon usage similarities also appear to be independent of replicon topology.In Chapter 3, we also identify the bacterial and archaeal genomes that are the most heterogeneous and homogeneous in codon usage—a characteristic that can be assessed by determining the number of genes that are significantly different from the modal codon usage of the genome.We find that the genomes with the most homogeneous codon usage are predominantly from organisms with reduced genomes including endosymbionts, parasites, and free-living marine bacteria.The most heterogeneous genomes include members of the genera Bacteroides, Corynebacterium, Xylella, Neisseria, Bifidobacterium, and Desulfotaela.In these latter organisms, greater than 2/3 of the genes in the genome differ significantly from the mode.In Chapter 4, we provide a method for evaluating expression-related codon usage bias (a major source of heterogeneity within genomes).This method is based upon the calculation of an axis that intersects the modal codon usage of a genome and the mode of a set of highly expressed genes.We show that this method is well suited for evaluating expression-related codon usage bias in genomes with extreme base compositions, such as Pseudomonas aeruginosa (66% G+C for the genome), a problem that has plagued previous methods.This method also provides a criterion for identifying foreign genes that have been recently acquired by the genome via horizontal gene transfer.In Chapter 5, we use the mode to characterize the major codon usage groups within the genomes of Escherichia coli K-12 and Salmonella enterica subsp. enterica serovar Typhimurium LT2.When we compare the codon usages in these genomes, we find that the genes that have been recently acquired via horizontal gene transfer are more similar in codon usage than are the genes that have been vertically inherited.To explore the generality of this observation, we compare genomes of three Agrobacterium species and find that the modal codon usages of the plasmids from different species are more similar than the modal codon usages of the corresponding chromosomes.Implications of the methods and data presented in this dissertation, particularly their implications for the study of horizontal gene transfer, are discussed.