Patterns Of Nucleotides Flanking Substitutions In Bacterial Genome

BackgroundPatterns of nucleotide substitutions depend on the base composition of context sequence.The analysis of context-dependent effect helps to understand biological process, such asnucleotide substitution and DNA mismatch repair. Recent research focused oncontext-dependent effect in human and chloroplast genome. CpG methylation has beenstudied for almost30years. Previous study utilized a variety of experimental and theoreticalmethods, but often got inconsistent conclusion. Relative entropy, as research tools ininformation theory, could be used to interpret genetic information. In this paper, based onrelative entropy analysis and relative entropy contribution, we study the patterns of nucleotidesubstitutions’ flanking sites in bacterial genome.ResultsIn this study, relative entropy was used to study neighboring sites effect in20bacterialgenomes. The aligned complete genome sequences were analyzed to obtain the basecomposition of neighboring sites of substitutions, and to calculate relative entropies. Inbacterial genome, the relative entropies are often highest among all neighboring sites, andhigher than the threshold defined credibly. Similar to the study in human and mammaliangenome, three-base periodicity exists in flanking sites of some nucleotide substitutions. Toexplore the origin of three-base periodicity, we used the alignment data of ortholog genes inmouse and rat genome, and set stricter background sequences frequencies. With theelimination of synonymous codon usage bias, the three-base periodicity still exist in flankingsites, but are weaker than the patterns in human and mammalian genome.To explore the neighboring sites effect, the relative entropy contribution of each base wascalculated. The results show that TA dinucleotides tend to be substituted, and CpGmethylation previously found in mammalian genome also exists in bacterial genome.ConclusionsAccording to the analysis of sites adjacent to nucleotide substitutions, CpG methylationexists in chloroplast genome. The patterns flanking sites in closely related species are moresimilar than that in distantly related species. Our study further explains the origin ofthree-base periodicity. Synonymous codon usage bias is one of the most significant factors responsible for the observed three-base periodicity. The relative entropy method could bewidely used in sequence analysis in the near future.