NEW YORK (GenomeWeb News) – The yeast transcriptome is more complicated and convoluted than previously supposed, new research suggests.
Scientists from Yale University used a quantitative sequencing approach called RNA-Seq to map the transcriptome of the well-studied yeast model organism Saccharomyces cerevisiae, identifying coding regions as well as untranslated regions and introns. Nearly three quarters of the yeast genome was transcribed. And, the team reported, many transcripts differed from those described before. The results appeared online today in Science Express.
“It’s a good lesson,” senior author Michael Snyder, a molecular, cellular, and developmental biologist at Yale University, told GenomeWeb Daily News. “Even though it’s the best characterized genome, there’s still a lot to be improved upon.”
In the past, several approaches have been used to identify yeast transcripts, such as analyzing large open reading frames, looking for conserved sequences, or using cDNAs to probe tiling microarrays. But these do not capture all of the transcript information, Snyder said.
For example, for tiling microarrays, cross-hybridization of probes can give false positives, he explained. In contrast, the RNA-Seq approach has no cross-hybridization. In addition, Snyder said, it’s very quantitative, offers a larger dynamic range, and provides good boundary information.
For this paper, he and his team used the RNA-Seq approach to characterize the S. cerevisiae transcriptome. First, they isolated polyadenylated or poly(A) RNA from actively growing yeast cells and used reverse transcription, with two either random hexamers or oligo dinucleotide primers, to create cDNA. Then, they broke up the cDNA and sequenced the fragments using an Illumina Genome Analyzer.
When they mapped the reads — more than 14 million for hexamer primed samples and nearly 16 million for oligo(dT) primed samples — to the yeast genome, the researchers found that both approaches revealed similar transcription patterns.
Consistent with previous studies, the majority of the yeast genome — 74.5 percent — is transcribed. In particular, the team found high expression of genes involved in biosynthesis and ion transport.
Interestingly, their results predicted that 321 yeast gene transcripts have upstream ORFs within 5’ untranslated regions.
Based on their data, Snyder said, they also adjusted the 5’ annotation of several genes. For example, they found dozens of genes each for which the start codon may be further upstream or downstream than previously believed.
They also found some unanticipated features at the 3’ ends of genes. For instance, when the researchers looked at poly(A) sites, they found that 540 genes appear to have more than one poly(A) site. Another unexpected result: many yeast genes had overlapping 3’ ends. The authors speculated that these may represent some form of gene regulation.
“Surprisingly, we found evidence that the transcription of a large number of yeast genes overlaps with transcription from the other strand,” the authors wrote. “Pervasive overlapping transcripts may be unique to S. cerevisiae and other organisms lacking miRNA/siRNA processing components which might otherwise degrade double stranded RNAs.”
And while they verified 240 of 306 known introns, the team found at least four cases where previously annotated introns were expressed at the same levels as exons next door. For two of these, the inclusion or exclusion of introns would change the protein sequence predicted. This could be especially important given that so many people are working with yeast open reading frames and using genetic constructs in biochemical experiments, Snyder noted.
“[O]ur novel RNA-Seq method allowed us to map the transcriptional landscape of the yeast genome and define UTRs and novel transcribed regions,” the authors noted. In the future, Snyder said, the group plans to track the yeast transcriptome throughout the cell cycle and to expand their application of RNA-Seq to include other organisms.