NN-RNALoc: Neural network-based model for prediction of mRNA sub-cellular localization using distance-based...

Area of Science:

• Molecular Biology
• Bioinformatics
• Computational Biology

Background:

• Messenger RNA (mRNA) localization is vital for gene expression regulation and protein targeting.
• Aberrant mRNA localization is implicated in neuromuscular diseases and cancer.
• Accurate prediction of mRNA cellular location is essential for biological research.

Purpose of the Study:

• To introduce NN-RNALoc, a novel neural network-based method for predicting mRNA cellular localization.
• To develop efficient and accurate sequence representation for mRNA.
• To integrate protein-protein interaction data for improved prediction.

Main Methods:

• Developed a distance-based subsequence profile for efficient RNA sequence representation.
• Integrated protein-protein interaction data with novel mRNA sequence features.
• Utilized a neural network architecture for predicting mRNA localization.

Main Results:

• NN-RNALoc demonstrated superior accuracy compared to existing predictive models on benchmark datasets (CeFra-Seq and RNALocate).
• The method significantly reduced computation time compared to previous approaches.
• Achieved higher accuracy than established methods including mRNALoc, RNATracker, and DNN5-mer.

Conclusions:

• NN-RNALoc offers a more accurate and computationally efficient solution for predicting mRNA cellular localization.
• The integration of novel sequence features and protein interaction data enhances predictive performance.
• This tool has implications for understanding gene regulation and disease mechanisms related to RNA localization.