Mapping regulatory elements from transcription initation data

Determining enhancer and promoter activities is important for modeling gene regulation and interpreting the impact of genetic variation on disease. Studying regulatory element activity genome-wide has been challenging, partly due to a lack of benchmarks. Here, we systematically evaluate CAGE for detecting active regulatory elements and present PRIME, a framework for analyzing regulatory elements from transcription initiation data. We assess nuclear RNA CAGE (nucCAGE) versus total RNA CAGE using fine-mapped eQTLs, likely pathogenic variants, and CRISPRi loci. Transcribed loci identified by PRIME on CAGE data show superior performance in detecting active regulatory elements compared to transcription run-on assays, and nucCAGE improves recall. Importantly, nucCAGE maintains the ability of CAGE to study promoter activities at base-pair resolution. These assays and computational tools will aid in understanding gene regulatory circuits and interpreting pathogenic variants. Transcription initiation was profiled in 5 human cell lines (i.e., A549, HCT116, GM12878, HepG2, K562) using CAGE. Selected samples were subjected to 1) nuclear extraction by applying osmotic pressure and sheer forces or 2) transcriptional perturbation by applying triptolide. Libraries were prepared from defined numbers of input cells to allow comparability of results and sequenced to acquire strand-specific transcription start site signals to summarize promoter and enhancer activities.