UC San Diego

The identification of sequence elements responsible for transcriptional activity remains a difficult challenge in post-genomic biology. Advances in microarray and next- generation sequencing technology have increased the accuracy and resolution for determining co-regulated genes, the genomic localization of transcription factors, and the locations of transcriptional initiation. We have developed a computational framework, named HOMER, for the analysis of gene expression, ChIP-Seq, and RNA-Seq data that utilizes differential motif discovery to accurately identify transcription factor binding sites. We used HOMER to find the promoters of genes induced by different inflammatory stimuli, leading to the discovery of a novel cis-regulatory element, the type I interferon response element (T1ISRE), which resides in the promoters of genes specifically induced by type I interferons. We then applied these methods to embryonic stem cell factors and identified a nucleosome positioning pattern surrounding Nanog motifs that would prevent Oct4 and Sox2 binding if a nucleosome were present, suggesting a novel pioneering role for Nanog in dictating the accessibility of pluripotent enhancers. Analysis of high-throughput 5' RNA- Seq data indicated that only 20% of human and mouse and 40% of Drosophila promoters use a focused pattern of initiation and are enriched for position-specific core elements, such as the TATA box. Unexpectedly, nearly half of human and mouse promoters and 30% of Drosophila promoters appeared to contain nucleosome positioning sequences and rotationally constrained nucleosomes in phase with a selective group of upstream transcription factor binding sites. These features are associated with multiple sites of transcriptional initiation with a periodicity of approximately 10 bp and define a previously unrecognized class of 'periodic promoters'. Comparison of nucleosome positioning in human and Drosophila periodic promoters revealed a GC-dependent shift in sequence patterns that is predicted to facilitate a corresponding placement of the first downstream nucleosome relative to the sites of transcriptional initiation in both species. In contrast to focused promoters, which preferentially direct developmental and high-magnitude, signal-dependent programs of gene expression, periodic promoters are configured to preferentially direct constitutive expression of genes that support general cellular functions. Features of periodic promoters are evident in a wide range of metazoan species, suggesting an ancient origin for this initiation strategy