Raw signal segmentation for estimating RNA modification from Nanopore direct RNA sequencing data.

Area of Science:

• Genomics
• Bioinformatics
• Molecular Biology

Background:

• Accurate estimation of RNA modifications from direct RNA sequencing is crucial for biological research.
• Current computational methods struggle with raw signal segmentation, limiting RNA modification analysis.
• Direct RNA sequencing offers a powerful way to study RNA modifications.

Purpose of the Study:

• To develop an improved method for raw signal segmentation in direct RNA sequencing data.
• To enhance the accuracy of RNA modification site identification using Nanopore sequencing.
• To provide a more interpretable and robust computational tool for RNA analysis.

Main Methods:

• Developed SegPore, a novel computational method based on a molecular jiggling translocation hypothesis.
• Implemented SegPore as a white-box model to improve interpretability and reduce signal noise.
• Validated SegPore's performance against state-of-the-art methods like Nanopolish and Tombo on benchmark datasets.

Main Results:

• SegPore significantly outperforms Nanopolish and Tombo in raw signal segmentation accuracy.
• The SegPore method demonstrates superior performance in reducing structured noise within the raw signal.
• SegPore-enhanced analysis (SegPore+m6Anet) achieves state-of-the-art results in site-level m6A identification.

Conclusions:

• SegPore offers a substantial advancement in raw signal segmentation for direct RNA sequencing.
• Improved segmentation directly translates to enhanced accuracy in identifying RNA modifications, including m6A.
• SegPore provides a more reliable and interpretable tool for RNA modification studies.