CFPLncLoc: A multi-label lncRNA subcellular localization prediction based on Chaos game representation and...

Area of Science:

• Molecular Biology
• Bioinformatics
• Computational Biology

Background:

• Subcellular localization of long noncoding RNAs (lncRNAs) is crucial for understanding their biological functions.
• lncRNAs often exhibit multiple subcellular localizations, posing challenges for existing prediction methods.
• Current computational tools are limited in predicting multi-label lncRNA subcellular localization.

Purpose of the Study:

• To develop a novel deep learning model, CFPLncLoc, for predicting multi-label subcellular localization of lncRNAs.
• To leverage Chaos Game Representation (CGR) images of lncRNA sequences for improved prediction accuracy.
• To enhance feature representation and extraction for multi-label prediction tasks.

Main Methods:

• Utilized Chaos Game Representation (CGR) to convert lncRNA sequences into images.
• Implemented an Image Update Strategy (IUS) to enhance feature representation in CGR images.
• Integrated a Centralized Feature Pyramid (CFP) model from computer vision for multi-scale feature fusion.

Main Results:

• Ablation studies demonstrated that both IUS and CFP significantly improved prediction performance.
• CFPLncLoc outperformed existing state-of-the-art predictors on an independent test set.
• The model achieved superior results using the MaAUC evaluation metric.

Conclusions:

• CFPLncLoc provides an effective computational approach for predicting multi-label subcellular localization of lncRNAs.
• The integration of CGR, IUS, and CFP offers a powerful framework for lncRNA localization prediction.
• This tool advances the study of lncRNA functions by accurately identifying their subcellular distribution.