A revised annotation of the Helicobacter pylori genomes.

With the genomic era, huge amounts of information are generated. Therefore, biologists are increasingly demanding structured, exhaustive and comparative databases. The PyloriGene database (http://genolist.pasteur.fr/PyloriGene) was developed to answer to such challenges by integrating and connecting the information generated by the sequencing of two distinct strains of Helicobacter pylori, a human gastric pathogen. This pointed to the need for a general annotation consensus of both strains as their physical and functional annotation differed significantly in some cases. A revised functional classification was created in order to accommodate the existing data on both sequenced genomes and the possibility of classifying coding sequences (CDS) into several functional categories were used to harmonize CDS classification. Revision of the annotation of the two complete genomes in light of new data allowed us to reduce the percentage of hypothetical proteins from around 40% to 31%. This also resulted in a reassignment of functions or new functional assignments for a total of 108 CDS, representing around 7% of the total CDS. Interestingly, only around 5% of the total of CDS with an assigned function were based on work done on H. pylori. Finally, comparison of the two genomes revealed a significant amount of CDS size variation among corresponding (orthologues) CDS. Although a majority of these size variations were due to natural polymorphism, interestingly, as much as 113 CDS pairs differed in size due to different start codon assignment, a recurrent problem of genomes physical annotations. [ABSTRACT FROM AUTHOR]