LPI-SKF: Predicting lncRNA-Protein Interactions Using Similarity Kernel Fusions

Area of Science:

• Genomics and Molecular Biology
• Bioinformatics and Computational Biology

Background:

• Long non-coding RNAs (lncRNAs) are crucial regulators of diverse biological processes, including transcription, splicing, and translation.
• Understanding the functional mechanisms of lncRNAs heavily relies on studying their interactions with proteins.
• Accurate prediction of lncRNA-protein interactions is essential for advancing our knowledge of cellular regulation.

Purpose of the Study:

• To propose and evaluate a novel computational model, LPI-SKF, for predicting potential long non-coding RNA-protein interactions.
• To integrate diverse similarity measures of both lncRNAs and proteins to enhance prediction accuracy.
• To provide a tool applicable for identifying interactions involving novel or uncharacterized lncRNAs and proteins.

Main Methods:

• Development of the LPI-SKF model, integrating similarity kernel fusion (SKF) and Laplacian regularized least squares (LapRLS) algorithms.
• Incorporation of various similarity features for both lncRNAs and proteins within the LPI-SKF framework.
• Rigorous evaluation using 5-fold cross-validation to assess prediction performance.

Main Results:

• The LPI-SKF model achieved a high Area Under the Receiver Operating Curve (AUROC) of 0.909 in cross-validation.
• LPI-SKF demonstrated superior performance compared to existing state-of-the-art methods for lncRNA-protein interaction prediction.
• An impressive 95% (19 out of 20) of top-ranked predicted interactions were validated by existing biological data.

Conclusions:

• The LPI-SKF model is a powerful and accurate tool for predicting long non-coding RNA-protein interactions.
• LPI-SKF shows significant potential for discovering novel lncRNA-protein interactions, thereby advancing functional genomics research.
• The developed model and associated code are publicly available, facilitating further research and application.