Dynamic characterization and interpretation for protein-RNA interactions across diverse cellular conditions using...

Area of Science:

• Molecular Biology
• Bioinformatics
• Genomics

Background:

• RNA-binding proteins (RBPs) are critical regulators of gene expression.
• Understanding RNA-RBP interactions is key to deciphering RNA function in different cellular states.
• Existing computational methods struggle with cross-predicting RNA-protein binding across diverse cell lines and tissues.

Purpose of the Study:

• To develop an advanced computational framework for predicting dynamic RNA-protein binding events.
• To enhance the prediction accuracy of RNA-protein interactions across varied cellular conditions.
• To provide novel insights into RNA-RBP roles in pathological mechanisms and gene-disease associations.

Main Methods:

• Development of HDRNet, an end-to-end deep learning framework.
• Training and validation on 261 linear RNA datasets from eCLIP and CLIP-seq experiments.
• Incorporation of supplementary tissue-specific data for enhanced prediction.
• Application of motif and interpretation analyses for mechanistic insights.

Main Results:

• HDRNet demonstrates high accuracy and efficiency in identifying RNA-protein binding sites, especially for dynamic predictions.
• The framework outperforms existing state-of-the-art models on diverse RNA datasets.
• Motif and interpretation analyses reveal new perspectives on RNA-RBP interactions in disease.
• Functional genomic analysis highlights previously unrecognized gene-disease associations.

Conclusions:

• HDRNet offers a powerful tool for precisely predicting dynamic RNA-protein binding events.
• The study provides valuable insights into the molecular mechanisms of RNA-RBP interactions in health and disease.
• HDRNet facilitates the exploration of gene-disease relationships, advancing our understanding of genetic disorders.