Open solutions for pooled direct tRNA and rRNA sequencing, demultiplexing, and analysis | LC26

Abstract
Nanopore direct RNA sequencing (DRS) enables single-molecule analysis of RNA modifications. Originally designed for polyadenylated RNAs, community-driven protocol adaptations now allow sequencing of non-polyadenylated RNAs such as tRNAs and rRNAs. However, the absence of open (de)multiplexing solutions hampers large-scale studies and broader adoption of the technology. Multiplexing lowers sequencing costs by enabling parallel analysis of multiple samples using unique barcode sequences. Together with the von Kleist Lab, we have recently introduced WarpDemuX-tRNA, which is specifically optimized for nanopore tRNA sequencing. Generally, analysis of tRNA profiling data is complicated by their short length, sequence redundancy, and dense modification patterns. Here, we present QutRNA2, a scalable workflow that integrates graphics processing unit (GPU)-accelerated local alignment, statistical filtering, pairwise error profile comparison, and flexible visualization. QutRNA2 achieves up to 25-fold speed improvements over central processing unit (CPU)-based approaches and robustly detects enzyme-dependent modifications in both nuclear- and mitochondrial-encoded tRNAs, as demonstrated in large human and mouse datasets. Our open-source, multiplexing-compatible framework fills a critical gap in current tRNA analysis tools. Finally, and building on our published work, we introduce an improved sequence-specific nanopore rRNA sequencing protocol that enhances 28S rRNA capture and increases read length. Using this optimized protocol, we further extended the WarpDemuX toolbox by developing a (de)multiplexing strategy for sequence-specific DRS, termed WarpDemuX-rRNA. Although designed for rRNA sequencing, WarpDemuX-rRNA is readily transferable to other sequence-specific DRS applications, including the sequencing of non-polyadenylated RNA viruses. We demonstrate the performance of this approach using human cell lines with knockouts of distinct small nucleolar RNAs (snoRNAs) required for specific 2'-O-ribose methylations of 28S rRNA.

Biography
Christoph Dieterich received his PhD degree in Bioinformatics from FU Berlin/Max Planck Institute (MPI) in 2005. He is a full professor of Bioinformatics and Systems Cardiology at Heidelberg University and head of the Klaus Tschira Institute for Integrative Computational Cardiology. His research focuses on computational RNA biology with a particular emphasis on RNA modifications, splicing, and translation.