Abstracts

Genetic Bases of Biodiversity

Description of regulatory interactions not only is an important part of bacterial annotation per se, but also provides additional data for functional annotation of proteins. Indeed, in many cases the sequence similarity alone allows one to predict only a general biochemical function of a protein, and in such cases the fact that the corresponding gene belongs to a particular regulon is a strong indication of the cellular role of the protein.

Unfortunately, in most cases the reliability of rules for prediction of regulatory sites, such as consensus patterns, positional weight matrices, is unacceptably low. Indeed, if the recognition threshold is set so as not to lose any true site, the number of genes with candidate in upstream regions can reach 10% of the genome. It is clear that such non-specific predictions are of very limited usefulness. However, availability of complete genomes of related bacteria makes it possible to increase the specificity of predictions. Indeed, one can assume that the sets of co-regulated genes (regulons) are conserved in related genomes. This yields the consistency check for analysis of several genomes: true sites occur upstream of orthologous genes, whereas false positives are scattered at random.

This technique was applied to the analysis of several bacterial and archaeal regulons, in particular, sugar utilization, aromatic amino acid metabolism, multidrug resistance, and iron uptake regulons of enteric bacteria, heat shock and purine metabolism regulons of bacteria and archaea, arginine regulons and several other regulatory systems.

Variants of the basic technique can be used to analyze closely related genomes with considerable conservation of non-coding regions, metabolic systems subject to horizontal transfer, RNA regulatory structures, and the systems with no experimentally mapped sites.

Note. Abstracts are published in author's edition

|Home Page| |English Part|

Go to Home