GLOBAL ANALYSIS OF GENOMIC TEXTS - THE DISTRIBUTION OF AGCT TETRANUCLEOTIDES IN THE ESCHERICHIA-COLI AND BACILLUS-SUBTILIS GENOMES PREDICTSTRANSLATIONAL FRAMESHIFTING AND RIBOSOMAL HOPPING IN SEVERAL GENES

Present availability of the genomic text of bacteria allows assignment of biological known functions to many genes (typically, half of the g enome's gene content). It is now time to try and predict new unexpecte d functions, using inductive procedures that allow correlating the con tent of the genomic text to possible biological functions. We show her e that analysis of the genomes of Escherichia coli and Bacillus subtil is for the distribution of AGCT motifs predicts that genes exist for w hich the mRNA molecule can be translated as several different proteins synthesized after ribosomal frameshifting or hopping. Among these gen es we found that several coded for the same function in E. coli and B. subtilis. We analyzed in depth the situation of the infB gene (experi mentally known to specify synthesis of several proteins differing in t heir translation starts), the aceF/pdhC gene, the eno gene, and the rp lI gene. In addition, genes specific to E. coli were also studied: omp A, ompF and tolA (predicting epigenetic variation that could help esca pe infection by phages or colicins).