LncMamba: A deep learning model for LncRNA localization prediction based on the Mamba model.

Area of Science:

• Computational Biology
• Genomics
• Bioinformatics

Background:

• Predicting long non-coding RNA (LncRNA) subcellular localization is vital for understanding their diverse biological roles.
• Existing methods may lack the precision needed for complex LncRNA localization prediction.

Purpose of the Study:

• To develop a novel deep learning framework, LncMamba, for accurate LncRNA subcellular localization prediction.
• To explore the potential relationship between nucleotide motifs and LncRNA localization.

Main Methods:

• Proposed LncMamba, a deep learning framework incorporating a two-layer Feature Pyramid Network (FPN) for multi-scale feature extraction.
• Introduced the Mamba network and an improved localization-specific attention mechanism for enhanced motif focus.
• Performed statistical analysis on localization motifs.

Main Results:

• LncMamba demonstrates a novel approach to LncRNA localization prediction using advanced deep learning architectures.
• The improved attention mechanism effectively identifies key sequence motifs relevant to subcellular localization.
• Statistical analysis suggests a link between specific nucleotide motifs and LncRNA localization patterns.

Conclusions:

• LncMamba offers a powerful new tool for predicting LncRNA subcellular localization, advancing the field of non-coding RNA research.
• The findings highlight the importance of sequence motifs in determining LncRNA localization and function.
• This work provides a foundation for further investigation into the regulatory roles of LncRNAs.