#10 · Tools & infrastructure

Sequence context biases long-read RNA sequencing and Pol II disassociation mapping

The common thread is that sequence context shapes what sequencing measurements report, and computational models are used to separate that bias from biology. In direct RNA nanopore reads, non-random error hotspots are linked to RNA folding after translocation, which can bias modification detection on CY1 viral RNA. In nascent RNA-seq, a probabilistic mixture model identifies two conserved Pol II disassociation region types, T-rich and GC-rich, that respond differently to stress, and base composition is reported to influence the position and precision of disassociation. A long-read PacBio pipeline extends this theme to viral vector genomes, tiling reads and applying a formal grammar to classify single AAV vector genomes into structural classes. What is new here is the explicit attribution of systematic nanopore errors to post-translocation folding, together with grammar-based single-molecule structural classification of AAV genomes.