KD-KLNMF: Identification of lncRNAs subcellular localization with multiple features and nonnegative matrix...

Area of Science:

• Genomics
• Bioinformatics
• Molecular Biology

Background:

• Long non-coding RNAs (lncRNAs) are functional RNA molecules >200 nucleotides.
• lncRNA subcellular localization is crucial for understanding their biological functions.
• Accurate prediction of lncRNA localization is essential.

Purpose of the Study:

• To develop a novel computational model for predicting lncRNA subcellular localization.
• To improve the accuracy and efficiency of lncRNA localization identification.
• To provide a feasible tool for researchers in the field.

Main Methods:

• Incorporation of k-mer and dinucleotide-based spatial autocorrelation for feature extraction.
• Application of Synthetic Minority Over-sampling Technique to address imbalanced datasets.
• Utilizing Kullback-Leibler divergence-based nonnegative matrix factorization for feature selection.
• Employing Support Vector Machine as the final classifier.

Main Results:

• The KD-KLNMF model achieved 97.24% accuracy on the training dataset.
• The model demonstrated 92.86% accuracy on an independent dataset.
• Performance surpassed previous methods for lncRNA subcellular localization prediction.

Conclusions:

• The KD-KLNMF model is a highly accurate and feasible tool for predicting lncRNA subcellular localization.
• The findings contribute to a better understanding of lncRNA functions.
• The developed model and source code are publicly available for research use.