RNAincoder: a deep learning-based encoder for RNA and RNA-associated interaction.

Area of Science:

• Computational Biology
• Bioinformatics
• Molecular Biology

Background:

• Ribonucleic acids (RNAs) are crucial in biological processes, interacting with proteins, compounds, and other RNAs.
• Predicting these interactions computationally is vital but relies on effective data representation.
• Existing methods often struggle to integrate both interacting partners into a single, computer-readable format.

Purpose of the Study:

• To develop a novel computational tool, RNAincoder, for encoding RNA-associated interactions.
• To provide a comprehensive set of RNA encoding features and a deep learning-based strategy for interaction representation.
• To enable efficient scanning of feature combinations for optimal prediction performance.

Main Methods:

• RNAincoder utilizes a deep learning-based embedding strategy for interaction representation.
• It incorporates a comprehensive collection of RNA encoding features.
• The tool performs large-scale scanning of feature combinations to identify optimal predictive models.

Main Results:

• RNAincoder effectively represents RNA-associated interactions.
• Its performance was validated using benchmark datasets and case studies.
• The tool demonstrated superior accuracy in representing RNA-associated interactions compared to existing methods.

Conclusions:

• RNAincoder offers an accurate and indispensable method for representing RNA-associated interactions.
• It complements existing computational tools for RNA interaction prediction.
• The tool facilitates a deeper understanding of RNA's role in physiological and pathological processes.