DeepmRNALoc: A Novel Predictor of Eukaryotic mRNA Subcellular Localization Based on Deep Learning.

Area of Science:

• Molecular Biology
• Bioinformatics
• Computational Biology

Background:

• Subcellular localization of messenger RNA (mRNA) dictates protein synthesis sites and functions.
• Experimental determination of mRNA localization is resource-intensive and slow.
• Existing computational methods for mRNA localization prediction require enhancement.

Purpose of the Study:

• To develop an advanced computational method for predicting eukaryotic mRNA subcellular localization.
• To improve the accuracy and efficiency of mRNA localization prediction compared to existing algorithms.

Main Methods:

• A deep neural network (DNN) architecture named DeepmRNALoc was designed.
• A two-stage feature extraction strategy was employed, involving bimodal information splitting/fusing and a VGGNet-like Convolutional Neural Network (CNN) module.
• The model was evaluated using five-fold cross-validation.

Main Results:

• DeepmRNALoc achieved high prediction accuracies for specific subcellular locations: cytoplasm (0.895), mitochondria (0.944), and nucleus (0.865).
• The method demonstrated superior performance over existing models and techniques in predicting mRNA subcellular localization.
• Moderate accuracies were observed for endoplasmic reticulum (0.594) and extracellular region (0.308), indicating areas for future refinement.

Conclusions:

• DeepmRNALoc represents a significant advancement in predicting eukaryotic mRNA subcellular localization using deep learning.
• The proposed method offers a more efficient and accurate alternative to experimental approaches.
• Further research may focus on optimizing performance for less accurately predicted locations.