SAGE identifies novel transcripts in Drosophila

Read Also

Genes corresponding to SAGE tags were identified using a Drosophila tag-to-gene mapping computer program (E. Pleasance, M. Marra and  S. Jones, submitted).  To facilitate the mapping of our salivary gland SAGE tags, as well as aid in gene annotation and gene discovery in Drosophila, we incorporated into the program sequence from 5,181 high quality salivary gland 3’ ESTs representing 1,696 different transcripts (Supplementary Material).  2,866 (61.9%) of the 4,628 different SAGE tags were mapped to known or predicted Drosophila genes, 289 tags (6.2%) were mapped to genomic DNA and ESTs not associated with a predicted gene, 1170 (25.3%) were mapped to genomic DNA and/or the reverse strand of an EST or predicted gene, and 303 (6.5%) could not be matched to existing sequence data.  The 303 unmapped tags could be due to sequence polymorphisms, common sequencing errors, or to lack of representation in the available sequence resources (e.g. heterochromatic regions).  It is also possible that unmapped tags span adjacent exons that are currently not represented in the EST or cDNA data set.

Our tag-to-gene mapping results highlight the main advantage of the SAGE method compared to other expression profiling methods such as oligonucleotide or cDNA array based analyses.  SAGE has the potential to reveal transcripts not previously identified.  More than 25% of the tags mapped only to genomic DNA and may represent novel genes or novel 3’ ends or splice forms of already predicted genes.  In at least 225 cases, our data suggests the existence of previously unpredicted transcripts that likely represent divergently transcribed overlapping gene sequences.  Our detection of these 225 transcripts is not surprising because current gene finding programs are unable to readily detect overlapping genes [8].  A complete list of salivary gland SAGE tag sequences, frequency, and mappings can be found in Supplementary Material Table S1.

            To validate our SAGE data, we conducted real-time quantitative RT-PCR analyses to verify differential expression of individual genes (Supplementary Material), and we searched for SAGE tags corresponding to genes associated previously with salivary gland cell death [3-6].  We detected expression of the ecdysone-induced primary response genes, E74, E75 and E93, and the cell death genes ark, dronc, crq, rpr and iap2 (see Supplementary Material for details).  In general, the gene expression profiles generated by SAGE were consistent with previous reports and could temporally distinguish known upstream ecdysone-induced transcriptional regulators from downstream death effector molecules.