RNA Modification Detection Using Nanopore Direct RNA Sequencing and nanoDoc2

Area of Science:

• Molecular Biology
• Bioinformatics
• Genomics

Background:

• RNA modifications are crucial for diverse cellular processes, including splicing, translation, and decay.
• Direct RNA sequencing using Oxford Nanopore Technologies (ONT) offers a powerful method for studying RNA modifications.
• Analyzing the complex raw signal data from ONT presents significant bioinformatics challenges for accurate modification detection.

Purpose of the Study:

• To introduce nanoDoc2, a novel software designed for detecting multiple RNA modifications from ONT direct RNA sequencing data.
• To address the informatics challenges associated with analyzing complex raw signal data from ONT.
• To provide a robust tool for precise identification of RNA modifications.

Main Methods:

• nanoDoc2 employs a signal segmentation algorithm utilizing trace value, a base probability feature from ONT's Guppy basecalling.
• A machine learning approach, specifically deep one-class classification with a WaveNet-based neural network, analyzes segmented raw current signals (6-mer).
• The software outputs a statistical score for each candidate position to indicate detected RNA modifications.

Main Results:

• nanoDoc2 successfully detects multiple types of RNA modifications from nanopore direct RNA sequencing data.
• The software provides a statistical score for precise localization of RNA modifications.
• Detailed instructions for using nanoDoc2, including signal segmentation, neural network training/testing, and prediction, are described.

Conclusions:

• nanoDoc2 offers an effective solution for the accurate detection of RNA modifications using ONT direct RNA sequencing.
• The developed software overcomes key bioinformatics challenges in analyzing complex nanopore sequencing data.
• nanoDoc2 facilitates deeper insights into the functional roles of RNA modifications through precise detection.