Deciphering RNA modification and post-transcriptional regulation with NetRNApan.

Area of Science:

• Molecular Biology
• Bioinformatics
• Genomics

Background:

• RNA modifications are vital for biological functions and disease.
• Existing prediction algorithms lack interpretability and generalizability.
• Discovering novel post-transcriptional regulations remains a challenge.

Purpose of the Study:

• Introduce NetRNApan, a deep learning framework for RNA modification site prediction.
• Enable motif discovery and trans-regulatory factor identification.
• Enhance understanding of RNA modification functions and mRNA regulation.

Main Methods:

• Developed a deep learning framework, NetRNApan.
• Utilized m5U profiles from FICC-seq and miCLIP-seq technologies.
• Applied NetRNApan to m6A sites from multiple experiments.

Main Results:

• NetRNApan demonstrated high accuracy and interpretability in RNA modification prediction.
• Identified five clusters with consensus motifs for m5U modification.
• Discovered 21 potential RNA-binding proteins (RBPs) linked to m5U modification motifs.

Conclusions:

• NetRNApan offers an accurate, interpretable, and generalizable approach for RNA modification studies.
• The framework provides insights into RNA modification functions and mRNA regulation.
• NetRNApan facilitates the discovery of novel regulatory mechanisms in RNA biology.