Absolute quantification and profiling of cellular tRNAs by nanopore sequencing and a deep learning classifier.

An end-to-end deep learning nano tRNA classifier achieves 98.2% tRNA classification accuracy from nanopore direct RNA sequencing ionic currents across all Escherichia coli tRNA species.

• Why it matters: tRNA abundance and modification states are dynamically regulated and affect translation efficiency, mRNA stability, and proteostasis, but sequencing-based profiling is biased by modifications.
• What they did: The model classifies tRNAs directly from nanopore-DRS ionic current signals without basecalling, and uses an equimolar native tRNA mixture as a ground-truth reference to correct read-count bias.
• The result: Bias correction enables absolute quantification and profiling under diverse Escherichia coli culture conditions, and binary classifiers detect specific tRNA modifications at the single-molecule level with high accuracy.

AI summary of the abstract · may contain errors · not medical advice · read the paper.