The Replication Strand Bias And Its Underlying Mechanism In Bacteria

There exist some deviations of biological characteristics between leading and lagging strand which caused by asymmetric replication mode.The strand bias has found its way in increasing organisms with the advent ofsoaring sequencing data, clarifying the difference betweentwo replication strands is of great significance to understanding of replication mechanism. However, so far no research focuses on various strand bias and their internal mechanism systematically. Thus, this research fully explores the replication strand bias in bacteria. The whole process and results in this paper can be divided into three parts:1. The first part specializes in the replication strand biases which are widespread in bacterial genomes. Here, we discuss 5 types of biases, including gene orientation, the number of open reading frames(ORF), nucleotide composition, substitution rate, and gene length between leading and lagging strands. For each type of strand bias, related studies and underlying mechanisms(mutation and/or selection are discussed) were summarized and Clostridium acetobutylicum ATCC 824 was used as a representative example to illustrate corresponding bias. Our results in this bacteriaindicate that there is little asymmetry between 2 replication strands on open reading frame number and gene length, whereas the other 3 features presented significant strand bias.2. Because the nucleotide composition is the basis for genomic research, composition decides structure, so we concentrate on the nucleotide composition bias and its contributing factor in this part.Our result indicates that:(1) the extent of bias for obligate intracellular bacteria is stronger than that for other species as a whole.(2) The Fusobacteria owned the highest mean bias value among all 19 phyla, and the Firmicutes comes second.(3) Strength of selected codon usage bias and generation time are both not observably related to composition bias.(4) There exists significantly negative relationship between composition bias and GC content, genome size, rearrangement frequency, COG functional subcategories A, C, I, Q separately.(5) Gene density, COG functional subcategories D, F, J, L, and V are positively connected with composition bias.(6) Among the various genomic features which are obviously relevant to composition bias, the gene density of leading strand makes the largest contribution. Therefore, strand composition bias is influenced by multiple factors with varying weights.3. Since in some bacteria, the intensity of strand composition bias was so strong that genes on the two replicating strands could be separated on the basis of their codon usage. This phenomenon is widely found in various organisms, which has important biological significances. Aiming to this, a online open database about codon usage-CCD(Codon Correspondence Analysis Database)was established. The codon usages were studied by using CA(correspondence analysis) and WCA(within-group correspondence analysis)for 267 species separately. Results show that codon usage bias exists between genes resided on leading and lagging strands in 59 species. Here, we take the species Borrelia duttonii Ly as an example to introduce the usage of CCD database. The CCD database is freely available at http://cobi.uestc.edu.cn/services/ccd/.This paper makes detailed discussion of replication strand bias in bacteria from transverse and longitudinal aspects. The analysis on thousands of species will provide us with a comprehensive understanding on these problems.