Development of quantitative RNA sequencing methods to understand RNA variant diversity.

Abstract

The advent of next-generation sequencing (NGS) has accelerated biological research by providing the opportunity to connect genome level sequence information to function. One key application of NGS has been analysis of gene expression and variation by implementing ways to recapitulate RNA level differences accurately. However, for studies that require accurate estimates of closely related variants or isoforms, as in case of evolutionary dynamics of viral quasispecies, each of the available sequencing platforms have critical limitations.

In this thesis work, we have developed methods to eliminate three of the significant limitations of current NGS platforms. These new methods tackle (i) Inability to reconstruct individual whole viral genomes due to genome fragmentation during sequence library preparations in short-read platforms (ii) High sequencing errors associated with 1D Oxford Nanopore sequencing platform and (iii) High background levels of host RNA hampering the detection of pathogen or novel RNA species in metagenomics studies. Our three novel workflows, namely, Single RNA sequencing (sRNA-Seq), TelN proteolemerase based nanopore 2D sequencing (T2D-NanoSeq), and Direct RNA amplification (D-RAMP), are designed to resolve each of these three limitations respectively. We have demonstrated each of the workflows, benchmarked them using synthetic samples and measured their efficiency.

Our first method, sRNA-Seq, can report sequences distantly separated on a single RNA molecule with > 10% recovery and 87% specificity. On the other hand, T2D-NanoSeq, has 40% efficiency in covalently linking both forward and reverse sequences of a duplex DNA that enables tandem reads of both strands which can allow us to reduce sequencing errors. Finally, our third method can generate short DNA probes against target (host) RNA molecules from a physical sample of the RNA and has demonstrated a ~5 fold decrease in the target RNA compared to non-targeted RNA. Overall, these approaches provide improved sequencing alternatives for quantitative and accurate RNA sequencing.