Platform for transcriptome-wide RNA modification identification in long reads

点击翻译按钮获取中文摘要

Project Summary RNA modifications are pervasive throughout the human transcriptome and affect transcript stability, localization, and function. In particular, ADAR-mediated adenosine-to-inosine (A-to-I) edits in RNA have been shown to affect pre-mRNA splicing and alter codon sequence. Amino acid changes caused by inosines have been implicated in various deleterious conditions, which include cancer and diseases of the brain. However, previous literature mapping inosine positions in high-throughput were only able to do so inside the limited context provided by short RNA-Seq reads. As modifications can be transcript-specific, elucidating the association of inosines with full mRNA isoforms is crucial for a more rigorous understanding of the role of inosine modifications in the tissues of our body and, more broadly, disease. Therefore, I propose to investiate A-to-I editing in the context of diseased and non-diseased systems using full-length mRNA nanopore sequencing. The nanopore is able to sequence whole RNA strands by converting changes in electrical current caused by RNA translocating through the pore into nucleotide sequence. ​Aim 1 leverages high-accuracy nanopore cDNA sequencing of cellular systems with and without ADAR knockdown to interrogate ADAR function and A-to-I-induced changes to transcript expression changes. To accomplish the latter, I will develop workflows to determine isoform structure from noisy, long reads. In addition to sequencing full-length transcripts, nanopore native RNA (nvRNA) sequencing informs on RNA modifications, as modified nucleotides appear as subtle alteration in current signal with respect to canonical nucleotides. As such, ​Aim 2 ​employs a generalizable approach to producing cost-effective training data for systematically understanding how inosines alter current signals in nanopores. I will use a Cas13b-ADAR fusion protein (REPAIRv2) to create site-specific edits and then perform nvRNA sequencing on the edited transcriptome. Site-specific A-to-I editing allows this approach to create a labelled inosine dataset in nvRNA signal from which I can develop computational algorithms to reliably identify inosines in nvRNA data. The REPAIRv2 approach to can be generalized to eventually identify any RNA modification with nanopores. ​Aim 3 will elucidate how A-to-I editing differs between tissues. I will sequence 4 normal tissue types with nvRNA sequencing, generating a map of A-to-I edits in conjunction with isoform usage using the software I am developing. Taken together, the fulfillment of these aims will not only provide further insights on elusive ADAR mechanism, but also create workflows for nanopore data analysis and a platform for the study of any modification. As I work toward my Ph.D. with this interdisciplinary project, I will gain invaluable skills in experimental and computational biology that will prepare me for a career in science.