Back

tRNA isodecoder analysis using Nanopore ionic current signals and deep learning

Akeson, S.; Kakhaki, P. D.; Esfahani, N. G.; Reinsch, J. L.; Barry, M. L.; Zamecnik, M.; Tzadikario, T.; Abu-Shumays, R. L.; Garcia, D. M.; Koutmou, K.; Jain, M.

2025-12-28 genomics
10.64898/2025.12.27.696718 bioRxiv
Show abstract

We describe application of Nanopore direct tRNA sequencing and ionic current signal for isodecoder-level analysis of bacterial and yeast tRNA. The computational advance is the application of deep learning for predicting tRNA isodecoders directly from raw nanopore ionic current signals. We then combined ionic current analysis with a pairwise sequence alignment strategy to improve tRNA sequence alignments. We applied this strategy to Nanopore data for E. coli and S. cerevisiae tRNA and aligned 2.6% and 13.1% more reads respectively when compared to present strategies for Nanopore tRNA analysis. Notably, this increase in alignments also came with a significant increase in alignment identity. The ionic current models confirmed enrichment of both an individual tRNA isotype and a tRNA fraction in separate experiments. Finally, we demonstrate that raw ionic current analysis is information-rich and has the potential to deconvolute dense chemical modifications in complex molecules like tRNA. This is an advance for tRNA analysis in prokaryotes and lower eukaryotes with potential to elucidate the role of tRNA in discovery biology and in human health applications. SIMPLIFIED BANNER GRAPHIC

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.