Elucidating metabolic pathways and digging for genes of unknown function in microbial communities: the riboswitch approach.

In the current post-genomic era, only 3% of all genes have been annotated based on experimental evidence. Even though functions can readily be predicted for many genes, 25% of these are likely to be wrong. The most widely used methods for function prediction rely on sequence similarity, which might be misleading in many cases. Other methods such as genomic context or phylogenetic profiles have been developed to increase gene annotation accuracy; nevertheless these are only efficient when complete genome sequences are available. Here we propose a new approach based on riboswitch identification. Riboswitches are highly conserved regulators of gene expression located in the 5' untranslated region of certain genes. When transcribed they adopt three-dimensional structures that recognize their ligands with great affinity and specificity. This specificity is a key issue for our method, allowing functional assignment with great accuracy.