Computational and Experimental Modeling of Alternative Polyadenylation

项目摘要/摘要

Project Summary / Abstract Alternative polyadenylation (APA) plays an important role in the post-transcriptional regulation of most human genes. The broad importance of APA is well exemplified by the altered expression of NUDT21, a key APA regulator that we reported in the first cycle of this grant, in diseases such as glioblastoma, diopathic pulmonary fibrosis, and neuropsychiatric disorders. More recently, our work has revealed a novel mechanism by which 3ʹ- UTR shortening can repress tumor suppressor genes (e.g., PTEN) in trans by disrupting competing endogenous RNA (ceRNA) crosstalk, rather than by inducing oncogenes in cis. Aside from these few examples, the prevalence and functions of APA in a wide spectrum of human traits and diseases remain largely unknown. Most human traits/diseases have been found as associated with hundreds of thousands of noncoding single-nucleotide polymorphisms (SNPs) in numerous genome-wide association studies (GWASs). However, functional interpretation of these SNPs remains a significant challenge because GWAS data do not show how the SNPs work. To better understand their effects, expression quantitative trait loci (eQTLs) have been widely used to link GWAS SNPs to gene expression. Despite massive efforts on elucidating eQTLs, the functions of many GWAS SNPs remain unexplained. An important reason is that eQTLs do not consider APA regulation. We have recently constructed the first human 3′UTR APA quantitative trait loci (3′aQTLs), which contain ~0.4 million SNPs associated with APA of target genes, using ~8,000 GTEx v7 RNA-seq samples across 46 tissue types (Nature Genetics, accepted in 2021). These 3′aQTLs can explain ~16.1% of GWAS SNPs in 15 common traits/diseases, and they are largely distinct from eQTLs and splicing QTLs. Based on these exciting preliminary data, we hypothesize that computational and experimental modeling of APA will substantially facilitate the interpretation of numerous GWAS SNPs important for APA regulation, which are enriched in 3′UTRs and gene downstream regions. Hence, we propose to develop innovative bioinformatics and experimental methods for identifying 3′aQTLs and nominating APA-linked disease/trait susceptibility genes in a wide variety of cell types and environmental factors, followed by in vivo functional characterization using our unique CRISPR engineering system. We expect to establish APA as an emerging and important molecular phenotype to explain a large fraction of GWAS risk SNPs, leading to significant novel biological insights into the genetic basis of APA and APA-linked susceptibility genes in a wide spectrum of human traits and diseases.