Project description:Most of human genome may show evidence of transcription, yet annotated transcripts account for less than 5%. The basis for this major discrepancy is not clear, and it remains uncertain whether excess transcription is functional, or simply a byproduct of pervasive, non-specific RNA polymerase binding and transcription initiation. Understanding the default state of the genome would be informative in determining whether the observed pervasive activity is functional. The genome of any extant organism has undergone billions of years of evolution, making it unclear whether any observed genomic activity, or lack thereof, has been selected for. We sought to address this question by introducing a completely novel 101 kb locus into the genomes of two eukaryotic organisms, S. cerevisiae and M. musculus, and characterizing its genomic activity based on chromatin accessibility, chromatin states, and transcription. The locus was designed by reversing, but not complementing, the sequence of the human HPRT1 locus, including ~30 kb of both upstream and downstream regulatory regions, allowing retention of sequence features like repeat frequency and GC content but ablating coding information and transcription factor binding sites. We also compared this reversed locus with a synthetic version of the normal human HPRT1 locus in both organismal contexts. Neither the synthetic HPRT1 locus nor its reversed version evolved to harbor yeast promoters. Nevertheless, we observed widespread transcriptional activity of both loci in yeast, and this pervasive activity was observed both when the loci were present as episomes and when chromosomally integrated. In the latter case, it was obvious that the accessibility and level of transcription initiation substantially exceeded that of the flanking native yeast genome sequences. In contrast, when integrated in the mouse genome, the synthetic HPRT1 locus showed transcriptional activity corresponding precisely to the HPRT1 coding sequence, whereas the reverse locus displayed no activity at all, but was instead actively repressed by Polycomb machinery. Together, these results show that novel genomic sequences lacking coding information are active in yeast, but repressed in mouse stem cells, indicating a major difference in default genomic states between these two divergent eukaryotes, with implications for understanding pervasive transcription and the birth of new genes.

Dataset Information

Synthetic STARR-seq reveals how DNA shape and sequence modulate transcriptional output and noise.

Publications

Synthetic STARR-seq reveals how DNA shape and sequence modulate transcriptional output and noise.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets