Comparative genomics study of inverted repeats in bacteria

We investigate the number of inverted repeats observed in 37 complete genomes of bacteria. The number of inverted repeats observed is much higher than expected using Markovian models of DNA sequences in most of the eubacteria. By using the information annotated in the genomes we discover that in most of the eubacteria the inverted repeats of stem length longer than 8 nucleotides preferentially locate near the 3' end of the nearest coding regions. We also show that IRs characterized by large values of the stem length locate preferentially in short non-coding regions bounded by two 3' ends of convergent genes. By using the program TransTerm recently introduced to predict transcription terminators in bacterial genomes, we conclude that only a part of the observed inverted repeats fullfills the model requirements characterizing rho-independent termination in several genomes.