CSAIL Event Calendar

Modeling Cellular Transcription: How the Cell Finds its Genes

Transcript definition is central to the expression of genes and the functional organization of genomic sequence. Yet, in general, we do not know what signals eukaryotic cells use to define individual genes. Here, we have derived remarkably simple classifiers that can identify yeast promoters, terminators, and transcript bodies and rely solely on nucleic acid sequence features detectable by the cell through DNA- and RNA-binding proteins. For promoters, we show that the predicted promoter-defining cis-elements are sufficient to drive transcription, and that mutation of the trans-acting factors responsible for this definition results in predictable expression changes genome-wide. For terminators, we show that Hrp1 binding sites are the major determinant of where cleavage and polyadenylation will occur and that cleavage sites are frequently bidirectional. We developed a hidden Markov model for unifying these three classifiers into a single model that predicts transcript structure and identifies initiation and termination sites more accurately than the individual models, indicating that these features influence each others’ function via their relative positions. Our model also provides an explanation for non-genic transcripts, which are predicted to arise from bidirectional promoters and promoter-like sequences and are stabilized by the transcription of terminator-like sequences. Our model represents the first concrete description how transcriptional units are defined by the cell on a genome-wide scale and provides insight into the evolution of the genome.

See other events happening in January 2013