A reporter assay (CapSTARR-seq) was performed in the GM12878 cell line, in order to generate a new dataset - in addition to the ones publicly available - on which to predict Epromoters. We selected 28 reporter assay datasets to find candidate promoters which show enhancer function, i.e. a comprehensive resource of candidate Epromoters, in 11 cell lines. We further characterized the candidate Eprom for epigenomic markers, showing that they are distinct from a set of control promoters. Then, we overlapped the candidate Epromoter regions with over 180,000 SNPs that are associated with diverse traits and diseases, yielding 4,330 GWAS-SNPs in 2,301 Epromoters, showing that Epromoters are in fact enriched for variants that are associated with more GWAS traits that even fall into different categories, indicating the potential pleiotropic effect of Epromoters. Here, pleiotropy is defined as a single element affecting more than one trait independently. Using eQTL data, we show that the candidate variant Epromoters are in fact associated with the altered expression of multiple target genes. Additionally, using massive parallel reporter assay data, we find that the GWAS-SNPs with validated allelic impact are present more often in Epromoters (with higher pleiotropy) than in control promoters, further validating the potential pleiotropic function of Epromoters. We experimentally validated one of these SNPs associated with COVID19 risq. These SNP lied within the OAS3 Epromoter and we showed that genetic variation impact on both the promoter and enhancer activity of this Epromoter.
These results beg the question, how do variants influence the enhancer versus promoter function of Epromoters? A number of candidate variant Epromoters were selected for genome editing using CRISPR/Cas9, and later prime editing to introduce the minor and major alleles to determine the effect of the variants on enhancer versus promoter function and thus the transcriptional changes induced in the target genes. Unfortunately, no clones were obtained that incorporated the selected variants into the genome, despite thorough efforts with multiple techniques and cell lines. Therefore, we pursued a different, more high-throughput approach; MPRA. Using the same synthesized sequences of the two alleles of variants in candidate Epromoters, we cloned the sequences in two orientations in a vector to test separately enhancer or promoter activity. Encountering a number of obstacles to data interpretation along the way, we are currently analyzing the data.
