Genomic detection of new yeast pre-mRNA 3 '-end-processing signals

To investigate Saccharomyces cerevisiae 3'-end-processing signals, a set of 1352 unique pre-mRNA 3'-end-processing sites, corresponding to 861 differe nt genes, was identified by alignment of expressed sequence tag sequences w ith the complete yeast genome. Nucleotide word frequencies in the vicinity of the cleavage sites were analyzed to reveal the signal element features. In addition to previously recognized processing signals, two previously unc haracterized components of the 3'-end-processing signal sequence were disco vered, specifically a predominance of U-rich sequences located on either si de of the cleavage site. One of these, the downstream U-rich signal, provid es a further link between the 3'-end-processing mechanisms of yeast and hig her eukaryotes. Analysis of the complete set of 3'-end-processing sites by means of a discrimination function supports a 'contextual' model in which t he sum total effectiveness of the signals in all four elements determines w hether or not processing occurs.