ModiCal: A Targeted Calibration Workflow for Site-Specific m⁵C Validation by Nanopore Direct RNA Sequencing

Area of Science:

• Molecular Biology
• Bioinformatics
• Genomics

Background:

• Accurate identification of RNA 5-methylcytidine (m⁵C) at single-nucleotide resolution is challenging for nanopore direct RNA sequencing (DRS).
• Existing methods often have high false-positive rates and lack site-specific accuracy for transcriptome-wide profiling.

Purpose of the Study:

• To repurpose the ModiDeC classifier into a targeted, high-precision validation tool for RNA modification sites.
• To establish a reproducible and transferable framework for reliable site-specific m⁵C confirmation using DRS.

Main Methods:

• A three-step calibration workflow integrating biochemical and computational modules, using synthetic RNAs and yeast models.
• Iterative retraining of the neural network with unmodified RNA signals to eliminate false positives.
• Application to human prerRNA and dengue virus genomic RNA for validation across diverse contexts.

Main Results:

• The refined ModiDeC accurately detected known m⁵C sites and eliminated off-target predictions.
• The method successfully identified previously undetected m⁵C sites and resolved distinct deposition regimes.
• The calibration logic demonstrated transferability across different RNA types.

Conclusions:

• A robust framework combining biochemical validation and neural network refinement enables reliable site-specific m⁵C detection via DRS.
• This approach significantly improves the accuracy and precision of RNA modification analysis.
• The developed method is transferable and applicable to various RNA contexts for future research.