UTR reconstruction and analysis using genomically aligned EST sequences

Untranslated regions (UTR) play important roles in the post-transcriptional regulation of mRNA processing. There is a wealth of UTR-related information to be mined from the rapidly accumulating EST collections. A computational tool, UTR-extender, has been developed to infer UTR sequences from genomically aligned ESTs. It can completely and accurately reconstruct 72 % of the 3 ’ UTRs and 15 % of the 5’ UTRs when tested using 908 functionally cloned transcripts. In addition, it predicts extensions for 11 % of the 5 ’ UTRs and 28 % of the 3 ’ UTRs. These extension regions are validated by examining splicing frequencies and conservation levels. We also developed a method called polyadenylation site scan (PASS) to precisely map polyadenylation sites in human genomic sequences. A PASS analysis of 908 genic regions estimates that 40-50 % of human genes undergo alternative polyadenylation. Using EST redundancy to assess expression levels, we also find that genes with short 3 ’ UTRs tend to be highly expressed